ABSTRACT. 1. The research question Central America faces inclusive development challenges of great magnitude. Inclusiveness is understood as to include the elimination of poverty, inequality, and underpinning the rights of people to a reasonable level of material well-being and social and citizenship rights. This ample vision of inclusive development not only has a positive impact on social well-being but also drives an economic system that is more propitious for learning, innovation, and higher productivity. Government policies consider inclusive development as a priority, but progress in their application is slow due in part to the variation in how policies are implemented. Aguirre-Bastos et al. (2014) attribute this problem to the lack of common theoretical and empirical grounding, whereby most countries operate their social programs in relative isolation from other policies. Sutz and Tomasini (2013) point out that isolated policies are not sufficient to face the challenges of inclusive development. These authors in analyzing an international organization study on poverty alleviation note “claiming centrality for social policies without empowering them through their alliance with other public policies will bear little effect”, and that “it is striking the sort of invisibility that knowledge, science, technology, and innovation have for many of those that fight against poverty and inequality”. Underpinned by such context, and to overcome the practice of rhetoric declarations on the value of innovation for inclusion, but little effective action, a key research question can be set as to how can the gaps in the innovation-inclusion nexus be closed in the Central American region? Against this backdrop, since 2023, a regional research and development Project in Central America, is being conducted by the Central American Higher University Council with the financial support of the International Development Research Center of Canada. 2. Methodology In early 2023, the Project called for projects to concentrate on Indigenous and rural territories where poverty, inequality, and lack of social rights are evident. Once identified, project leaders would approach local communities to request their participation in the definition of activities that allow, on one hand, to facilitate the introduction of new technologies to their productive systems and facilitate environmental management, and on the other hand, define future actions to empower the community in not only managing such technologies, but also defining their future development process by proposing and adopting territorial research and innovation policies. The call emphasized that gender equality should be stressed in the proposals, considering that overcoming gaps in gender conditions is fundamental for an inclusive development. Further to support the Central American integration process, proposals were expected to include institutions of at least three different countries of the region, thus creating collaborative networks as instruments to enhance this process. A total de sixteen projects were accepted for execution for a period of 22 months starting around September 2023. 3. Findings After a year and a half of execution, projects already show important outputs to determine the most significant variables for community development through socio-technological innovation, social inclusion, gender equity, and environmental management. To this end, projects have designed and implemented participative diagnosis to identify the development conditions of vulnerable populations in Indigenous and rural territories (Outputs of papers and reports are found at: https://csuca.org/es/idrc-csuca/) Information was collected through field visits and focus groups that included municipal authorities, women leaders, and farmers. Problems to allow understanding of local conditions and tailoring project strategies adjusted to the specific needs of the communities were discussed. Some projects searched for strengthening collaborative research and social innovation capabilities by integrating traditional knowledge under an interdisciplinary and gender approach. Several projects have dealt with various aspects of food security by strengthening agri-food risk management and developing human and technological capacities to establish agroecological production systems contributing to security and are resilient to climate change in the Dry Corridor of Guatemala, El Salvador, and Honduras. One project is developing distinct types of foods with functional characteristics from the use of local products as an alternative to food and nutritional security. The Project has collected raw materials and characterized bioactive molecules using pressurized fluid and High-Performance Liquid Chromatography-Mass Spectrometry to develop functional foods from local ingredients. Also, information aimed at community leaders was provided to familiarize them with the use of new food products derived from such local ingredients. One other important output is the establishment of a Central American network of biodiversity observers and researchers to propose solutions to conservation problems through data science. A tool has been developed and applied to collect data for needs expressed by the community. The projects are also contributing to add value to the region's academic/scientific forums related to the Intergovernmental Panel on Climate Change. Activities along this line are facilitating the creation of a large network of Central American researchers (and the region’s diaspora) working on climate change. One project is now establishing a network of Central American researchers in high-performance computing and its socio-environmental applications to promote the scientific inclusion of women and other vulnerable groups. Activities included a High-Performance Computing School and Scientific Computing Seminars. 4. Conclusion The first outcomes of the CSUCA – IDRC Project under execution foresee ways to close the innovation – inclusion gaps and approach public policy makers with sufficient evidence to enrich policies and define roadmaps with larger investments in human resources and regional research and innovation capacities. Such outputs will also contribute greatly to much-needed regional integration. References Aguirre-Bastos, C., Matos, S., Silvestro, B., Hall, J. (2014). Shaping National Innovation Systems for Inclusive Growth in Latin America: Towards a Research Agenda. Paper presented to the PRME (Principles for Responsible Management Education) Working Group on Poverty Conference. 28 – 30 July 2014, INCAE, Nicaragua Sutz, J., Tomasini, C. (2013) Knowledge, innovation, social inclusion, and their elusive articulation: when isolated policies are not enough. Conference: Paper presented at: International workshop on 'New Models of Innovation for Development, 4 and 5 July 2013, Manchester University.

ABSTRACT. How is responsibility practiced and made legible in in collaborations between universities and industry in the context of new emerging biotechnologies such as engineering biology? Engineering biology is the use of biological systems, such as microorganisms, cells, or enzymes, to make products like medicines, biofuels, food ingredients, and materials. It leverages biotechnology to grow and optimise these systems for large-scale production in industry, aiming to contribute to sustainable bioeconomies through biomanufacturing (Aguilar, 2019; Ayrapetyan et al. 2022). In multiple countries programmes have been initiated to develop and coordinate academic biotechnology research and industrial capabilities to generate scaled innovation in key sectors of pharmaceuticals, value-added chemicals, engineering materials, and advanced synthetic fuels.

Biomanufacturing also raises concerns about such issues as sustainability, resource inequality, biosecurity, dual-use, and public trust (Kemp et al, 2020; Shapira et al. 2022), which for example could include questions of dual use. For example, proponents emphasize their commitments to sustainability, claiming that biomanufacturing can provide alternatives to fossil fuel dependency. From this perspective, sustainability and commercialisation are imagined as working hand in hand. However, literature in responsible innovation questions this assumption, suggesting that sustainability and commercialisation are not inherently allied and require explicit and ongoing attention, reflection, engagement, and ethical scrutiny (Asveld 2021) and this is also relevant for biomanufacturing (Holland et al. 2024).

Moreover, biomanufacturing as part of the bioeconomy represents a powerful sociotechnical imaginary. It involves collections of (industrial and public) stakeholders, strengthened by the institutions that they inhabit, who promote specific visions about the benefits of commercial biotechnology that are powered by assumptions about how society should be organised. The concept of a bioeconomy can be understood as a sociotechnical imaginary, defined by Sheila Jasanoff (2015, p.4) as “collectively held, institutionally stabilized, and publicly performed visions of desirable futures, animated by shared understandings of forms of social life and social order attainable through, and supportive of, advances in science and technology”. Sociotechnical imaginaries analytically draw together the values and meanings embedded in innovations with their material manifestations through sociomaterial networks and assemblages. They pay close attention to questions of power, particularly where competing visions of the future struggle for dominance. Because of their relation to power is important to probe how these visions are co-produced by those working in academia and industry, particularly as they relate to understandings of how this can be done responsibly and in the public interest.

This paper derives from a study of a national UK biomanufacturing intermediary facility. This facility is established at a research university as a public-private partnership and designed to advance biomanufacturing commercialization. The facility is committed to innovating responsibly with a particular focus on sustainability. The overarching research question is: “how is responsibility practiced and made legible in the context of biomanufacturing in collaborations between universities and industry?” Sub research questions include, “how does responsibility as imagined by the biomanufacturing facility compare with public perspectives on commercial innovation?” and “how can industrial biomanufacturing be made publicly accountable?”

The research involves interviews and workshops with researchers and industry partners that seek to understand how responsibility in commercial biotechnology is understood and practiced. This empirical work further aims to uncover dilemmas, tensions, and contradictions alongside forms of legitimation and authority that are used to support specific views on responsibility. The research also involves analysis of grey literature associated with the facility and its research focus to get a sense of how responsibility has been institutionally encoded through discourses on governance, EDI, sustainability, and responsible innovation. The conference paper presents preliminary results from this research.

The research is undertaken within a framework that seeks to bring critical social science perspectives to innovation policy and practices. This involves exploring issues of power, governance, responsibility, and public accountability. These issues are the core focus of the field of responsible innovation (RI), understood broadly as “taking care of the future through collective stewardship of science and innovation in the present” (Stilgoe, Owen, Macnaghten 2013). RI is a field of academic study and a framework of ethical practice, making it highly suitable for a study of this kind which seeks to both understand and advise on responsibility. RI views responsibility less as a series of principles or values that are adhered to and more as a context-specific set of practices that involves anticipating future impacts of innovation, reflecting on underlying values and assumptions driving research, including diverse viewpoints on how innovation and knowledge should be implemented, and responding and adapting to these insights in ways that align innovation with broad societal needs. From this perspective, innovation is responsible when it is publicly accountable (Ribeiro et al. 2018), while also adhering to the more narrowly defined criteria of research ethics and integrity, and legal and regulatory requirements. The research iteratively analyses the data both inductively and deductively – working from the data but guided by the epistemic and political framing provided by RI and sociotechnical imaginaries. The data is analysed alongside findings from previous research on public views of what it means for innovation to be responsible to understand how responsibility is co-produced and how it compares to public understandings.

Prior work has pointed out that scientists’ and industry professionals’ ideas about which issues are relevant to ethical consideration do not align with the publics’ (Macnaghten, Kearnes, Wynne 2005). Where science and industry typically focus on risk categories that can be quantified and measured, the public is more concerned about issues of control, governance, purposes, and ownership (Macnaghten, Kearnes, Wynne 2005). Seen from this angle, the strategies for sustainability and commercialisation viewed by engineering biology research and industry stakeholders as inherently responsible might simultaneously be perceived by publics as a fundamental ethical tension. The research aims to offer fresh insights on this conundrum, probing how biomanufacturing imagining interacts with public understandings of responsibility and exploring opportunities for policy and institutional strategies that can address this tension.

(References omitted in this abstract)

ABSTRACT. The pursuit of economic progress, social inclusion and environmental sustainability in contemporary societies in the Global South necessitates a nuanced understanding of local innovation governance capacity (Kuhlmann and Ordoñez-Matamoros, 2017; Pérez and Ordóñez-Matamoros; 2022), specially when it comes to better understanding its role in fostering structural transformation via both technological and social innovation (Chataway et al, 2017; Ordoñez-Matamoros et al, 2021a; Pinzón-Camargo et. al., 2023).

As Schot and Steinmueller (2018) have pointed out, despite the growing recognition of social innovation as a catalyst for positive societal change, its conceptual ambiguity and the absence of robust governance and measurement frameworks (Van der Have and Rubalcaba, 2016), pose significant challenges to policymakers and practitioners alike (Edwards-Schachter, M. and Ordoñez-Matamoros, H.G., 2023). This paper examines the intricate interplay between local governance capacity and inclusive and sustainable innovation, with a focus on Global South contexts: paradoxical combinations of, on the one hand, poverty, inequality, unemployment, informal or even illegal ecosystems, institutional sclerosis, corruption, social injustice, insecurity, violence and/or environmental vulnerability, among others and, on the other hand, entrepreneurship, resilience, cultural diversity, resourcefulness and creativity of its population and institutions and/or abundant natural wealth, among other features. We use Colombian local communities as example.

Drawing on a multi-disciplinary literature, this study explores the evolving conceptual landscape of social innovation (Chataway et al, 2014; Fressoli et al, 2014; Papaioannou, 2014; Bergek et al, 2015), the role of social capital, trust and coordination as enablers of innovation (Pylypenko, H.M, Yu.I. et.al 2023), the role of insecurity and corruption within it, and their overall implications for local innovation governance structures in developing countries (Bortagaray and Ordoñez-Matamoros, 2012; Ordoñez-Matamoros, 2021b; Pinzón and Centeno, 2021; Pinzón, Ordoñez-Matamoros and Kuhlmann, 2022).

By employing a mixed research methodology, including policy and academic documents consulted, statistical and case studies analyses and stakeholder interviews, the paper studies the mechanisms through which local governance capacity depends on social capital structures instrumental for facilitating (or not) innovation initiatives aimed at promoting economic development, social inclusion, well-being and environmental sustainability in contrasting contexts, beyond innovation policies centered on fostering the productivity and competitiveness of enterprises.

We found common denominators worth exploring further with respect to the ways in which local governments go about encouraging inclusive and sustainable innovations in such a particular type of contexts, and on the ways in which such processes are being accounted for.

The findings underscore the complex dynamics inherent in local governance systems, highlighting the need for stakeholder network and adaptive and context- sensitive approaches to foster inclusive and sustainable innovation processes effectively, including sort out problems defined by communities, but deriving lessons worth exploring beyond the Colombian cases studied. Indeed, the study identifies key determinants of local governance capacity, including social capital based on trust and coordination systems, institutional arrangements, participatory mechanisms, innovation process led by communities themselves and resource allocation strategies, which significantly influence the emergence and diffusion of inclusive and sustainable innovations.

ABSTRACT. This study explores the nuanced impacts of tenure-specific human capital loss on firm innovation and examines the moderating role of female representation in mitigating these effects. Human capital has been considered as a critical driver for innovation. Firms increasingly grapple with the challenges posed by a competitive labor market and rising employee turnover. The increasing pressures have heightened concerns over the detrimental effects of human capital loss on innovation. In this context, researchers have emphasized the importance of preserving firm-specific knowledge to sustain innovation.

Despite extensive research on human capital loss, much of the focus has been limited to specific types, such as voluntary turnover, and narrow dimensions of its impact, particularly financial performance. The knowledge loss induced by employee turnover and its impact on innovation still remain underexplored. Furthermore, organizational tenure, which leads to varied levels of accumulated firm-specific knowledge, has been less highlighted. These gaps highlight the need for further exploration of the impact of employee turnover on knowledge retention and innovation, particularly focusing on organizational tenure as a key determinant of firm-specific knowledge preservation.

Building on the exploration of how the loss of experienced employees impacts human capital, it is highly necessary to explore how organizational strategies might lessen its impact on innovation. In particular, female representation within the workforce may hold untapped potential in mitigating these challenges. While female representation in enhancing firm performance has been studied, most research overlooks its broader workforce contributions and potential to mitigate the negative effects of human capital loss. This study addresses these gaps by drawing on the Knowledge-Based View and Human Capital Theory to further examine how tenure-specific human capital loss—classified as newcomers, tenured workers, and retirees—impacts firm innovation, and whether female representation serves as a contextual moderator in this relationship.

This study utilizes data from Korean firms collected in the 2011 (4th), 2013 (5th), 2015 (6th), and 2017 (7th) waves of the Human Capital Corporate Panel (HCCP). Our analysis focuses on 1,011 observations from Korean manufacturing firms. Firm innovation is measured by the number of patent registrations filed with the Korean Intellectual Property Office (KIPO), a widely recognized indicator of innovative output. Human capital loss is measured as the ratio of employees at different organizational tenure stages who left the company. Female representation is measured using two metrics: (1) the proportion of female employees relative to the total workforce and (2) a decile rank based on female workforce percentage. To test the hypotheses, this study applies a Poisson quasi-maximum likelihood (PQML) estimator to address the over-dispersed nature of patent count data while controlling for firm-level characteristics such as size, age, and workforce composition. Robustness checks include alternative modeling techniques, such as negative binomial regression, and supplemental measures of innovation (intellectual property) and female representation (Blau index).

Results reveal that human capital loss does not uniformly affect innovation. While general human capital loss has an insignificant impact, the loss of tenured workers significantly and negatively affects innovation. On the other hand, the loss of newcomers and retirees shows no significant effect. These findings align with the Knowledge-Based View and Human Capital Theory, which emphasize the critical importance of firm-specific tacit knowledge accumulated over time. Moreover, we found that female representation within the workforce plays a significant moderating role, mitigating the negative effects of tenured worker loss on innovation. Firms with higher female representation experience a less pronounced decline in innovation, highlighting the importance of female representation in knowledge loss environments resulting from human capital loss.

This study makes a significant contribution to the literature by highlighting the differential impacts of tenure-specific human capital loss on firm innovation, emphasizing the unique vulnerability posed by the departure of tenured workers. Existing research on human capital loss has largely overlooked the role of organizational tenure and its impact on innovation, instead focusing on specific departure types, such as voluntary turnover, and financial outcomes like sales and productivity. Our study addresses these gaps by highlighting the nuanced, negative effects of tenure-specific human capital loss on innovation, revealing that the departure of tenured employees, who often hold firm-specific tacit knowledge, has the most detrimental impact compared to newcomers or retirees.

Our study broadens the understanding of female representation’s role, by shifting the focus from diversity-driven perspectives to a knowledge-based lens, extending its strategic importance in the mitigating the effects of knowledge loss. While previous research has focused on the impact of female representation on firm performance, studies addressing how gender moderates the effects of human capital loss on innovation are scarce. Building on prior work, our findings reveal that female representation within the workforce can mitigate the negative impact of human capital loss on innovation, surpassing the positive reinforcement effects typically examined.

Moreover, our research highlights the extensive influence of female representation across all workforce levels, shifting the emphasis away from the traditional focus on women in top management roles. While previous research has focused on women in decision-making roles, such as CEOs, our findings shed light on the underexplored contributions of female employees in workforce for mitigating the negative effects of human capital loss on firm innovation, pointing to its strategic importance in fostering resilience and sustaining innovation.

This study offers practical implications that can inform organizational strategies and policies. It is wise for organizations to develop targeted retention strategies for tenured employees and invest in mechanisms to preserve their tacit knowledge. Simultaneously, promoting female representation can strengthen organizational resilience, creating a collaborative and knowledge-retentive environment essential for mitigating the impacts of knowledge loss.

ABSTRACT. Countries with efficient innovation systems have sound foundations for accelerated economic growth. Given the emerging importance of the global south economies, comprehending their innovation systems would be significant; therefore, unlike previous studies, this paper endeavors to measure the efficiency of the National Innovation Systems (NISs) of the global south countries to identify the system (in)efficiencies. Generally, countries adopt ‘best practices’ through policy learning to calibrate sound policy directions for economic development. Therefore, cross-country examination of NIS efficiencies might be crucial to identifying weaknesses and benchmarks. This research aims to answer the following questions: Are NISs in the global south countries performing efficiently? And how are countries characterized according to their innovation efficiencies? Which factors countries need to focus to enhance their economic and innovation development? Following Choi & Zo's (2019) and Alnafrah's (2021) research frameworks, a two-stage relational DEA with a small addition of variables in the Knowledge Production Process (KPP) and Knowledge Application Process (KAP) is employed. The two-stage DEA model has the advantage over the traditional one-stage DEA model in that it analyzes the internal operations of the innovation process. Knowledge Absorption (KA), having FDI inflows and high technology imports as an additional construct for the other input of KAP, is also included as a pivotal factor in fostering innovation activities in the context of developing countries. Data for inputs and outputs of KPP and KAP is extracted with two years lag for both processes to measure the effect of inputs on outputs. With a sample of 80 countries from the global south, this study tries to fill the research gap by analyzing NIS countries covering all income levels. Data for the selected variables are sourced from various sources, e.g., the Global Innovation Index, the World Bank, and UNESCO SIU. The selection of the sample countries from the Global South is based on the availability of data on input and output variables. Additionally, this study has an enlarged sample of countries representing all economic strata, as previous studies have ignored this crucial aspect, especially from the Global South perspective. Findings suggest that China has the most efficient NIS, followed by Singapore and Vietnam. The relative efficiency of the countries revealed a significant performance gap in KPP and KAP. Azerbaijan, Algeria, and Cambodia showed lower KPP scores than their higher KAP scores. The same pattern is observed in several other countries. Based on the DEA results, clustering analysis is performed to discern typologies of countries matching their respective innovation performance characteristics. Countries are classified into several groups based on system performance and income status similarities to devise targeted policy recommendations. Interestingly, most countries are found the knowledge application leaders compared to their respective knowledge production status, thus demonstrating the imbalance between the two crucial stages of NIS. Such classifications and categorizations help provide insights to policymakers to assess and evaluate their county’s comparative economic status and utilization of resources. The results can help provide valuable insights to policymakers regarding the structural functioning of NIS and highlight the weak areas in which strategic actions should be directed. Furthermore, research findings would be helpful for targeted policy measures to improve and promote NIS efficiency as well as design, formulate, and implement policies intended for the fair allocation of resources within the NIS perspective.

ABSTRACT. Quality Matters: Enhancing Global STI Tracking Through the Assessment of Indicators

Introduction

Global STI indicators, such as those used in STI rankings by the Global Innovation Index and other scoreboards, are used to quantify the scientific and innovative capabilities and productivity of countries. They are relied upon by researchers, policy makers, and news outlets to produce studies and messages about scientific achievement and power. However, the extent to which they enable a comprehensive description of the global STI landscape is determined by geographical coverage and data variability. In this paper we: 1. compile a set of 333 STI indicators used by 16 global STI rankings / scoreboards 2. classify the indicators according to a novel conceptual framework that merges the national innovation systems and the input-output frameworks. 3. assess the quality of global STI indicators using statistical measures to evaluate coverage, timeliness, and coherence. 4. make the full time-series dataset publicly available along with our added classifications and quality assessments. 5. suggest ways to use this dataset to incorporate assessment of indicator quality

Our research offers a novel, freely-available datasource, enabling users to create their own analyses and scoreboards. It also underscores the need for conscientious scoreboard production that highlights data quality issues, improving the interpretation of STI evaluation and tracking.

Background

Existing scoreboards discuss succinctly features of quality, completeness, or appropriateness of their indicators. A handful of studies assess STI scoreboards, such as the European Innovation Scoreboard and the Global Innovation Index. Adam (2014), for instance, assesses the quality of the EIS. He finds issues in relation to: lack of a clear framework, arbitrary indicator selection, dependence of scores on the availability of indicators, large variability of the indicators, abrupt changes in the values of indicators between editions of the scoreboard, and comparability. His findings expand on earlier studies (Schibany & Streicher, 2008; Grupp & Schubert, 2010; Archibugi, Denni & Filippetti, 2009). Beyond scoreboards, researchers have analysed indicator quality. Some examples include: Molas-Gallart et al. (2002) on third mission indicators, Wouters et al. (2019) on open science indicators and others such as Wickson and Carew (2014). However, there is a need for a better and more systematic analysis of data quality in STI scoreboards in relation to their coverage, timeliness, and coherence. In this paper we address this gap and offer a dataset of indicators for users to do their own scoreboard and quality evaluation.

Methodology We collected STI indicators from 16 global, regional, and local STI scoreboards, resulting in a list of 1,013 indicators. After deduplication, we identified 333 unique indicators from public data sources, focusing on those with data available for at least one country from 2010-2019. Next, the indicators were classified using a novel conceptual framework that merges national innovation systems (Lundvall 1992) and input-output frameworks (OECD 1963). This classification considered four dimensions: STI domain, logical input-output framework (enablers, inputs, linkages, outputs, impact), key actors in the innovation system, and STI activities. This approach allowed a detailed assessment of the indicators' coverage and relevance. Finally, the quality of the STI indicators was assessed using statistical metrics that capture both indicator-level and country-level dimensions. At the indicator level, metrics include the percentage of missing countries to measure geographical coverage, the percentage of outliers to identify potential anomalies, and the coefficient of variation (CV) to evaluate data dispersion across countries and over time. At the country level, quality is assessed by calculating the average percentage of missing indicators, measurement periodicity to evaluate temporal consistency, and the CV to reflect variability in indicator values over a 10-year period. These metrics provide a systematic and quantitative framework to evaluate the coverage, timeliness, and coherence of the data for STI analysis.

Below we show a summary of the quality indicators and their calculation:

INDICATOR LEVEL:

-Percentage of Missing Values (Coverage): This metric represents the proportion of countries where a specific indicator was not measured during the analysis period.

-Percentage of Outliers (Coherence): This metric indicates the proportion of anomalous values for a specific indicator across countries and years.

-Coefficient of Variation (CV) (Coherence): This metric measures the relative dispersion of values for a specific indicator over countries and years.

COUNTRY LEVEL:

-Percentage of Missing Values (Coverage): This metric shows the proportion of indicators not measured for a given country over the 10-year period.

-Measurement Complementary-Periodicity (Timeliness): This metric indicates the percentage of years in which an indicator was recorded for a country during the analysis period.

-Coefficient of Variation (CV) (Coherence): This metric assesses the consistency of values for all indicators measured in a given country over the 10-year period.

Findings

Firstly, our analysis employs a novel framework combining national innovation systems and input-output approaches to map the STI system. This revealed significant gaps, particularly in the representation of the informal economy and non-academic actors, which remain largely unmeasured.

Secondly, on average countries miss 43.7% of the 333 indicators and 4 years of data (37%) in the 10-year period. Missing data is higher in input, linkage, and output indicators than in enablers and impact.

Thirdly, while many indicators show low percentages of outliers (6.1%), the high average coefficient of variation (0.69) indicates significant data dispersion, complicating cross-country comparisons. However, temporal trends within countries remain stable (0.25), reflecting internal consistency over time.

Finally, robust indicators related to enablers and impacts, such as the UN-SDGs Red List Index (mean CV = 0.24, missing countries = 2%, mean outlier percentage = 3%), show high coverage and coherence. These can serve as benchmarks to improve other indicators, enhancing their reliability for global STI evaluation.

Discussion The study highlights significant disparities in the quality of global STI indicators, particularly in terms of coverage and coherence. The uneven coverage and variability in data may be due to diverse collection methodologies, complicating interpretation of STI performance. Despite these challenges, some indicators related to enablers and impacts show higher quality and can serve as benchmarks. The study emphasizes the need for more rigor and coverage in producing indicators to better reflect innovation systems’ performance and inform global STI policy.

ABSTRACT. This study focuses on the use and construction of composite indicators (CIs), which are statistical tools used to evaluate performance of innovation systems by combining multiple variables into a single, comparative metric, Cherchye et al. (2007) and Greco et al. (2019). Most prominent examples of CIs to evaluate innovation systems are the European Innovation Scoreboard of the European Commission and the Global Innovation Index by the World Intellectual Property Organization, Edquist et al. (2018). CIs are instrumental for policymakers, academics, and stakeholders because they condense complex datasets into comprehensible and actionable scores, helping to assess and rank performance across countries, regions, or institutions.

Purpose and Methodology of CIs to Evaluate Innovation Performance

The primary objective of using CIs is to aggregate information from several variables into a singular score or index, representing the utility or effectiveness of an alternative. Decision-makers (DMs) use this single measure to compare the performance of various alternatives. Organizations globally have leveraged CIs to rank observations based, enabling comparative assessments across many areas, Frudenberg (2003) and Bandura (2006, 2008, 2011).

However, the methodology involved in constructing CIs is complex and introduces challenges. The design process must address multiple methodological questions, including variable selection, normalization, weight assignment, and the choice of an aggregation function. The Organization for Economic Co-operation and Development (OECD), for instance, has established a "toolbox for constructors" with guidelines to avoid common pitfalls like non-transparent weighting and improper data aggregation, OECD (2008).

Challenges in CI Construction

A major issue in CI construction is the selection of an aggregation function and the weights assigned to each variable. The aggregation function determines how the various variables interact within the composite index. Many institutions opt for a linear arithmetic mean function, assuming that all variables can substitute each other perfectly, but this assumption has been criticized. For example, in the Global Innovation Index used for raking innovation systems, increasing the R&D expenditure in the business sector (BERD) can always substitute a reduction in the number of Employed ICT specialists (EICTS) in the same amount. This linear approach implies that a high score in one variable can perfectly offset a low score in another, which may not always reflect reality, Barbero et al. (2021).

Alternatively, other forms of aggregation functions that incorporate varying levels of substitutability, such as geometric mean or fixed proportions, can be applied. For instance, the fixed proportions (Leontief function) assumes that each variable contributes a necessary amount to the total utility, making it suitable when all variables are essential to the evaluation. The choice of an aggregation function is thus subjective and represents the preferences of the DMs. Furthermore, establishing weights for the variables is critical, as it defines each variable's importance in contributing to the overall score. The relative weights illustrate the trade-offs or importance between variables, which heavily influences the CI outcomes.

Subjectivity in Weights for Innovation Indicators and the DEA-BoD Approach

The subjective nature of selecting weights introduces potential biases, as DMs' preferences may not align with the realities or priorities of each alternative being evaluated. To address this, an alternative method known as Data Envelopment Analysis with Benefit-of-the-Doubt (DEA-BoD) is often employed, Cherchye et al. (2004). DEA-BoD is an unsupervised method that removes the necessity for predetermined weights, instead allowing each alternative to be evaluated on its most favorable terms. This approach involves using mathematical programming to maximize the performance score for each alternative by adjusting weights in a way that reflects each alternative’s strengths.

The DEA-BoD approach evaluates each alternative based on a constructed utility frontier, meaning that each alternative is compared relative to the best-performing peers. This approach creates a Pareto-efficient evaluation, which identifies alternatives that cannot be improved without diminishing the performance of at least one variable. DEA-BoD thus avoids biases that could result from a predetermined, universal weighting system.

Limitations of DEA-BoD

Despite its strengths, DEA-BoD has certain limitations. Its flexibility allows for zero weights to be assigned to some variables, which can lead to misrepresentations. For instance, an alternative may score highly by maximizing one variable at the expense of others, which may not accurately reflect a compressive innovation system. As an example, imagine the evaluation of a national innovation system where R&D expenditure in the business sector (BERD) is assigned a zero value because the alternative has the largest value in the number of Employed ICT specialists (EICTS).

Hybrid Methodology: Combining CI and DEA-BoD

To address the drawbacks of both CI and DEA-BoD, the study introduces a hybrid methodology that incorporates elements of both. This approach uses a standard utility function and a common set of weights from CI methodology, while also leveraging DEA-BoD to enable flexibility in weighting for each alternative. This combination provides a dual assessment: one subjective, based on a chosen utility function and standardized weights, and the other objective, through DEA-BoD’s Pareto-efficient analysis.

In this framework, two types of effectiveness measures are generated: overall utility effectiveness and technical utility effectiveness. Overall utility effectiveness reflects subjective preferences, while technical utility effectiveness is purely quantity-based, comparing each alternative against the optimal, or Pareto-efficient, benchmark. This distinction enables the identification of an “allocative” utility effectiveness component, which quantifies the difference in performance attributable to the chosen weights and preferences of the DMs (Pastor et al., 2022).

Application and Validation to the European Innovation Scoreboard

The study applies the proposed methodology to data from the European Innovation Scoreboard, which is used to rank countries based on innovation indicators. By employing different utility functions and weight configurations, we demonstrate how rankings can vary based on subjective or objective evaluations. This allows for a more nuanced understanding of each alternative’s performance relative to others under various weighting schemes.

The hybrid methodology’s advantage lies in its capacity to leverage the strengths of both CI and DEA-BoD, offering a balanced evaluation that recognizes both the DM’s preferences and the objective performance of alternatives. This helps to bridge the gap between subjective and objective assessments, making it useful for situations where impartiality and fairness are critical.

ABSTRACT. Background With the rapid expansion of Machine Learning (ML) and Artificial Intelligence (AI) technologies in the last decades, Human-Computer Collaboration (HCC) emerged as a promising avenue for research and development (R&D) activities. HCC is a paradigm of knowledge discovery wherein at least one human agent engages with a computing system characterized by varying degrees of automation, autonomy, or resilience (Valdés-Pérez, 1999). Scientists claim that HCC expanded significantly towards offering new possibilities in scientific discovery and exploration, facilitating the idealization of research hypotheses that transcend human limitations (Tschandl, P. et al., 2020). HCC has played a significant role in fields as bionanomaterials research, i.e., a domain in which scientists manipulate complex biosystems processes and dynamics at nanoscale. In that area, the convergence of expertise from physics, chemistry, computation, engineering, and biology permits the development of new 2D/3D bio-architected nanomaterials which can be assembled autonomously to hold a desired function and/or a responsive behaviour to external triggers and stimuli (Shelin and Meenakshi, 2023). Bionanomaterials research focused on the understanding of biologically synthesized and conjugated nanoparticles and its applications, especially in the development of advanced nanoparticle-based targeted drug delivery systems (Kozierok et. al., 2021). However, some scientists grapple with the dual challenge of ensuring the reliability, replication and human validation of HCC-based findings and large database, given the escalating complexity inherent in such systems (Xiong et. al., 2022). We still lack a comprehensive understanding of how the advancement and diffusion of HCC systems are affecting the relationship, conduct, processes, and the regimes of governance of expert knowledge production in emerging areas of convergence S&T – as well as what ethical questions and regulatory problems it raises that could compromise the future of AI in scientific work. This warrants robust empirical investigation addressing how researchers frame the relevance of HCC in a bionanomaterial laboratory, and under which ways that data could inform the production of ethics and responsible research guidelines, education policies for STEM capacity and research, and Science and Technology policymaking.

Aim This article aims to investigate scientists’ attitudes toward the relevance of human-computer collaborations in a bionanomaterials laboratory in the United States. The study results from a thematic analysis of nine interviews applied with scientists of a respected research laboratory located in the New York Metropolitan Area in the summer of 2022.

Methods This study applied an ethnographic work in a research laboratory as its main qualitative research approach. By immersing in the field, ethnographers aim to gather real-world evidence to understand the cultural phenomena within specific organizational contexts and shared social frameworks, employing various qualitative methods of data collection and analysis (Roller & Lavrakas, 2015). As supported by Atkinson, Coffey, and Delamont (2001), this research employed participant observation, interviews, and informal interactions in the laboratory to gather data on scientific practices, narratives and routines shared among researchers, machines, and their algorithms – in terms of how they active affect the social order and technopolitical configuration of expert knowledge in bionanomaterials. By adhering to these methodological approaches, the study seeks to contribute to the expanding body of knowledge on Studies on Expertise within the fields of Sociology, and Science, Technology and Society studies (Eyal and Medvets, 2023). A total of nine staff scientists were interviewed for this study. The interviews consisted of two stages: an initial formal semi-structured conversation lasting approximately 30 minutes conducted in their respective offices or meeting rooms, followed by a more informal conversation during a guided tour of their facilities – aiming at interacting with the facilities, posters, photographs, and many machines in phase of prototyping, calibration or well-settled. Interviewees were selected based on their experience, roles, and availability within the laboratory, ensuring a comprehensive understanding of the work developed in that research facility. Interview transcripts were generated by TRINT and thematically analyzed through the assistance of Atlas.ti.

Results The findings reveal some factors affecting experts' decision making on automating research on bionanomaterials: scientist’s experience in the field; reflections on the cost-benefit of automation investments; relevance of human validation of experiments; collective trust in autonomously generated molecules and materials by peers; and communicational issues experienced between researchers holding multiple disciplinary affiliations. In areas of disruptive innovation such as bionanomaterials research, attitudes toward the relevance of HCC in the lab relate to stakeholders’ capacity to convince and produce consensus about why it is relevant to automate experiments and biomanufacturing work. While common sense might assume that automation is always desirable as a mechanism to enhance objectivity, precision, resilience, and productivity, it does evolve through routinary negotiations in specific organizational settings. By incentivizing the questioning among less experienced researchers on why algorithms are needed, Professors and Lab leaders support the co-production of new channels of dialogue that, not rarely, lands in the pragmatic frame/rational approach to decide about the use of scarce resources in the lab. Professors prioritize and foster human validation and re-experimentation of data to guarantee replicability and controlled levels of reproducibility.

Discussion Attitudes toward the relevance of HCC in that bionanomaterials research facility involve a dynamic and multi-dimensional process that encompass both technical and non-technical arguments. In this case, scientists’ discourses on the ethics and objectivity of their research share room with concepts of relevance and pragmatism as crucial determinants of when automation makes sense. The debate on when is relevant to automate is then responsible to co-create laboratory’s technopolitical infrastructures and narratives. Curiously, conversations about potential uses of AI to accelerate scientific research, in fact, stimulate researchers to value, even more, well-trained human resources to judge the epistemic aspects of HCC and its pragmatic relevance in the lab environment. This resonates with current debates on the overestimated role of AI and other algorithm optimization technologies in radical innovation areas, indicating that we need to critically examine how the preservation of scientific objectivity, technical interventions and judgment on relevance of automation balance each other to stabilize emerging science and technology avoiding noise and community’s distrust.

ABSTRACT. Artificial Intelligence (AI) is not just transforming the way we work—it is fundamentally redefining the future of scientific inquiry and innovation. This transformation demands a paradigm shift in how we prepare the scientific workforce of tomorrow and upskill today’s researchers. Despite its profound implications, science policy has yet to fully address the urgent need for a comprehensive framework to meet this challenge. This presentation examines the multifaceted impacts of AI on scientific workforce development, with a critical focus on curriculum innovation, equitable access to AI-driven tools, and ethical considerations in integrating AI into research practices. Drawing on a meta-synthesis of research on the future of work, we analyze how AI-driven workflows accelerate skill obsolescence and demand a reimagined approach to workforce readiness. However, the integration of AI into science is not without risks. Left unchecked, AI has the potential to erode the integrity of scientific processes and compromise science's role as humanity's most trusted epistemic practice. To mitigate these risks, we argue for a dual approach: equipping the scientific workforce with skills to ethically and creatively collaborate with AI, and developing robust standards for the production, evaluation, and critique of scientific knowledge in the AI era. Our findings offer actionable recommendations for policymakers, educators, and industry leaders. These include strategies to foster inclusivity, ensure responsible conduct, and enhance creativity in an AI-augmented scientific enterprise. By addressing these challenges collectively, we can secure the legacy of science as a driver of innovation and societal progress in the 21st century.

ABSTRACT. Background The advent of advanced generative artificial intelligence (AI) technologies has significantly transformed the academic landscape, particularly in the realm of research funding proposals. Scholars are increasingly leveraging AI to enhance efficiency in proposal development, aiming to secure funding essential for advancing science and technology. This shift presents profound implications for the research funding ecosystem, potentially increasing the volume of submissions, intensifying competition, and influencing the selection process of high-potential research projects. Research Method This study develops an economic model to analyze the broad implications of AI-assisted proposal writing on research funding dynamics and the advancement of innovation. By integrating game-theoretic analysis and simulation, we explore how AI assistance affects the number and quality of proposals submitted, the level of competition among researchers, and the probability of funding agencies selecting proposals with significant potential for scientific breakthroughs. The model simulates a competitive research funding environment where researchers decide whether to utilize AI assistance in crafting their proposals. Key components of the model include researcher decision-making, proposal quality and innovation potential, competition dynamics, and reviewer constraints in the selection process. Researchers weigh the benefits of AI assistance—such as reduced time and effort—against potential drawbacks like the risk of homogenization or reviewer biases. The decision impacts their ability to submit multiple proposals and allocate time to other academic pursuits. The model also considers how AI might influence the substantive quality of proposals, the potential for increased competition due to higher submission volumes, and the capacity limitations of reviewers. Through simulation, the study examines various scenarios reflecting different levels of AI adoption among researchers and its impact on the funding landscape. Key variables include the AI adoption rate, reviewer capacity, and funding agency policies. We analyze how these factors interact to influence the overall dynamics of research funding competition and the advancement of science and technology. Key Findings The findings indicate that AI assistance significantly lowers the barriers to proposal submission, leading to a substantial increase in the number of proposals. This surge intensifies competition for limited funding resources, potentially disadvantaging researchers who lack access to AI tools or who choose not to use them. The increased volume of submissions can overwhelm funding agencies and reviewers, affecting the thoroughness and fairness of the evaluation process. While AI can enhance the presentation and clarity of proposals, there is a risk that over-reliance on AI may result in homogenization of content. Proposals may become similar in style and structure, making it challenging for truly innovative ideas to stand out. Researchers might invest less effort in developing novel concepts, relying instead on AI to embellish standard ideas. This could lead to a dilution of the overall quality of proposals, with originality and creativity being overshadowed by polished but less substantive submissions. The heightened competition and potential quality dilution might reduce the likelihood of funding agencies selecting proposals with significant potential for scientific breakthroughs. Reviewers may experience fatigue due to the increased workload, leading to superficial evaluations or reliance on heuristics that do not favor unconventional or groundbreaking proposals. Consequently, the advancement of science and technology could be hindered if the most promising research does not receive the necessary funding. Implications for Policy and Practice These findings have significant implications for policy and practice within the academic and funding communities. Funding agencies may need to establish clear guidelines on the acceptable use of AI in proposal development to maintain the integrity of the evaluation process. Implementing submission limits per researcher or introducing stricter preliminary screening processes could help manage the increased proposal volume without stifling innovation. Enhancing reviewer capacity is crucial to address the challenges posed by the surge in submissions. Investing in reviewer training and providing tools to handle increased workloads can improve the quality of evaluations. Strategies such as collaborative review models or leveraging AI for initial screenings might help mitigate reviewer fatigue and ensure more thorough assessments of each proposal's substantive merits. Promoting innovation remains a key concern. Funding agencies could introduce mechanisms to identify and prioritize high-potential research, such as dedicated programs for high-risk, high-reward projects or alternative evaluation metrics that value originality over presentation. Encouraging diversity in research ideas and methodologies can help ensure that groundbreaking proposals receive the attention they deserve, despite the competitive landscape. Ensuring equity among researchers is essential to prevent exacerbation of existing inequalities. Addressing disparities in access to AI tools—perhaps by providing resources or support for researchers in under-resourced institutions—can help maintain a level playing field. This approach ensures that all researchers have an equal opportunity to benefit from AI advancements without unfair disadvantages.

ABSTRACT. Drug development is a lengthy, complex, and costly process, requiring extensive clinical trials, substantial risks, and strict regulatory oversight. Artificial intelligence (AI) holds transformative potential across the entire drug development lifecycle—from accelerating drug discovery and optimizing clinical trials to enhancing manufacturing and supply chain logistics.

A key challenge for innovation policy is to ensure that spending on medicines yields the greatest societal return. While policies like price controls and weakened intellectual property (IP) protections may aim to lower drug costs, they can also dampen incentives for future research and development (R&D) without necessarily enhancing productivity. A more sustainable approach is to prioritize policies that maximize the value of R&D investment. Boosting R&D productivity, especially given declining returns and rising capital costs, offers a more effective path to improve public health outcomes, promote equitable access, and drive economic growth.

Technological advances, particularly AI, promise to enhance R&D productivity, accelerate access to innovative therapies, and improve health equity and competitiveness. By increasing drug output and fostering market competition, AI could expand access to novel therapies, delivering greater value for society. AI also holds potential to accelerate the discovery of therapeutic targets and drug candidates for diseases that currently lack effective treatments.

This report explores AI’s transformative role across various stages of drug development, presenting evidence on how AI can enhance drug discovery, diversify clinical trials, and optimize manufacturing. Additionally, it addresses challenges to AI adoption in drug development and provides policy recommendations for effective integration that safeguards patient safety.

Drug development complexity continues to grow due to advancing biological knowledge, stricter regulatory requirements, a focus on complex diseases like cancer, the need for diverse clinical trials, and the rise of precision medicine. AI can address many of these challenges by improving efficiency, identifying viable drug candidates faster, and reducing costly trial failures. Case studies in this report illustrate AI’s wide range of applications: improving inclusivity in clinical trials, leveraging electronic health records to identify underdiagnosed populations, accelerating drug discovery, and enhancing gene therapy production. By strengthening productivity, AI can support innovation, bolster economic resilience, and enhance global competitiveness. Furthermore, AI can streamline regulatory review, enabling faster and safer access to essential therapies while upholding rigorous safety standards.

A supportive public policy framework is essential for responsible AI adoption in drug development. Key policies include funding AI-related research, preparing the workforce for an AI-driven future, supporting privacy-preserving data sharing, fostering public-private partnerships, and developing risk-based regulatory approaches tailored to patient impact. Such policies are crucial to unlocking AI’s full potential in drug development, accelerating the availability of life-saving therapies, and ensuring broader access.

ABSTRACT. In the Anthropocene era, cities are grappling with the dual challenges of climate change adaptation and sustainable development. Rapid urbanization, historical development inequalities, and the escalating impacts of climate change exacerbate these challenges, particularly for underserved and marginalized urban communities in both the Global North and South. Such communities often bear disproportionate climate burdens, yet existing adaptation policies can inadvertently reinforce inequities. Given the complexities of urban systems and the potential policy trade-offs between adaptation and development, it is crucial to align adaptation efforts with development priorities to promote sustainable and resilient cities.

Achieving this alignment often requires innovations and system changes, with sustainability transitions serving as a key to reshaping urban futures. Through technological or institutional-level innovations, cities are encouraged to transform pressing challenges into attainable objectives. While recent studies have documented progress in such innovations, especially at the technological levels, the critical question remains: “innovation for what?” (Chaminade, 2020). Traditional approaches to innovation, often driven by private sector dynamism and risk-taking, are ill-equipped to address the systemic challenges posed by climate change. To confront these long-term sustainability issues, which often stem from entrenched inequities, smart, inclusive, and transformative innovations are necessary. As advocated by The Earth Institute at Columbia University and Ericsson (2016), “transformation of societies” beyond business-as-usual approaches is needed to achieve the Sustainable Development Goals (SDGs).

Thus, innovation for a sustainable future requires more than one-dimensional changes. To foster sustainable adaptation, it is essential to focus on vulnerability – specifically, as a critical intersection between adaptation and development. Although vulnerability is frequently highlighted in adaptation and development literature as a central consideration, most studies focus on assessing vulnerability as a set of localized indicators without fully examining its ontological role as the bridge between adaptation and development. This study aims to fill that gap by examining the “lock-ins” of vulnerability within urban adaptation planning. With this approach, the temporal and spatial dynamics of vulnerability can be captured in the feedback loop between adaptation and development.

This research introduces a theoretical framework centered on the “lock-ins of vulnerability,” a concept underexplored in prior studies despite its critical importance, as noted by the IPCC (2022), which has identified “increased evidence of maladaptation across many sectors and regions…that are difficult and expensive to change and exacerbate existing inequalities.” These lock-ins are path-dependent mechanisms rooted in the institutional, technological, and cultural aspects of adaptation planning that hinder cities’ progress toward resilience. By dissecting these mechanisms, this study sheds light on the underlying causes that exacerbate vulnerability, providing insights into transformative changes that may enable cities to shift from maladaptive practices to sustainable adaptation pathways.

To analyze this framework in urban practices, the study conducted a survey of urban practitioners from 42 cities and employed fuzzy-set Qualitative Comparative Analysis (fsQCA). This approach allowed for an in-depth investigation of how institutional, technological, and cultural lock-ins influence cities' sensitivity to climate-related disasters and their adaptive capacity. The fsQCA method uncovered different pathways that either reinforce or alleviate vulnerability, revealing the conditions under which lock-ins interact to prevent cities from achieving resilience. The results underscore the role of specific lock-ins in constraining cities’ ability to respond to climate challenges and demonstrate the ways in which these lock-ins intersect to trap urban systems in the vicious cycle of vulnerability.

Notably, the study found that institutional and cultural lock-ins play particularly significant roles in shaping adaptive capacity. Cities with higher adaptive capacities often share unlocked institutional and cultural barriers, which would otherwise limit their ability to implement effective adaptation strategies. In the meantime, high sensitivity to climate risks arises from multiple different conditions depending on different local contexts although often closely related to rigid institutional frameworks and technological constraints that resist change. The findings suggest that policy interventions aimed at enhancing adaptive capacity must address these institutional and cultural constraints. By breaking these lock-ins, cities can create more inclusive and flexible environments that support transformative adaptation efforts.

Further analysis revealed a strong association between institutional and cultural lock-ins, with the two factors often reinforcing one another. This interdependence was evident in cases where overcoming institutional barriers also led to shifts in cultural and cognitive norms. Specifically, breaking institutional lock-ins was sufficient to unlock cultural and behavioral conventions in certain contexts, indicating that institutional reforms can pave the way for broader societal change. This finding emphasizes the need for integrated policy approaches that simultaneously address multiple lock-in dimensions to foster more adaptive urban environments.

The study contributes valuable insights into the systemic barriers that hinder urban adaptation and provides empirical evidence to support targeted interventions for sustainable adaptation. By identifying the complex interactions among different lock-ins, this research offers a framework for city officials and practitioners to understand how specific constraints can be dismantled to promote resilience. Moreover, the study’s findings advocate for a shift toward “transformational adaptation” – an approach that entails comprehensive changes in adaptation planning, combining technological innovation, institutional reform, and cultural transformation. Such an approach encourages cities to question established assumptions and values, fostering an environment where adaptive capacity can be strengthened across diverse urban contexts.

In conclusion, this study underscores the need for cities to address the multi-dimensional lock-ins of vulnerability within adaptation planning. The lock-ins of vulnerability framework offers both theoretical insights and practical guidance for policymakers seeking to transform maladaptive practices into sustainable adaptation pathways. By focusing on transformational adaptation, cities can move away from the vicious cycle of vulnerability and toward the virtuous cycle of sustainable adaptation and development. This research provides a valuable resource for urban practitioners aiming to design solutions that not only enhance adaptive capacity but also ensure that adaptation efforts align with broader development goals, paving the way for resilient and sustainable urban futures.

ABSTRACT. Motivation: Improving environmental outcomes will depend, in part, on technological innovation. But innovation is not value-free; it reflects the diverse motivations and imaginations of innovators. Innovation happens in both the public and the private sector. Typically, the private sector is motivated by private value creation and the public sector is motivated by public value creation, guided by their institutional mission. For the latter, 'mission-oriented' innovation policy has recently gained traction as a promise for sustainability transitions. While there is no shortage of research on the public sector's contribution to innovation, much of this is based on the assumption that all innovation is good. As a result, these studies focus largely on 'how much' rather than 'what kinds' of innovations are publicly supported.

This research aims to understand the innovative behavior of public sector agencies in the context of their missions, with a particular focus on sustainability. Specifically, now that climate change and sustainability have been on the agendas of (some) public institutions for decades, the question is whether this mission narrative has tracked with innovative action. In short: To what extent are public actors innovating in line with their sustainability missions?

Empirical setting: I consider this question in the context of plant breeding innovation in the US. I focus on the time before and after the introduction of an environmental sustainability mission narrative in the US Department of Agriculture (USDA), initiated by the creation of the Sustainable Agriculture Research and Education program (1991-present).

Plant patent data: To analyze the choices of public innovators in plant breeding, I rely on plant patent (PP) data from 1930 to 2022. I analyze only agricultural crops, the targets of agricultural innovation for sustainability, narrowing the number of patents used in the analysis from 33,154 to 5,409. Patent data are available through various US Patent and Trademark Office sources, and a combination of computational methods were used to collect, clean, validate, and combine them. The patent data relevant to this analysis include the following: abstracts (summaries describing the invention and its claims to novelty), assignees (legal owners), assignee classification (sectors of owners, e.g. individual or government), and government interest (funding declaration or interest by supplemental conveyance).

Key variables and analysis: This paper analyzes whether public innovators are breeding in line with their mission (i.e. prioritizing sustainability-related traits in the 21st century). Public innovation is operationalized as patents that meet one of the following criteria: are owned or funded by the federal government ('government interest') or developed at public universities. Private innovation, on the other hand, is owned by companies or individuals. Mission alignment is determined by specific plant traits that promote sustainability. To operationalize plant traits, I classify patent abstracts based on their different claims to novelty for the plant. Along with two research assistants, I manually classified a sample of 112 abstracts (2% of population) into 10 themes. This classified sample was used to prompt and validate OpenAI's GPT 4.0 model, which classified the remaining abstracts. Model classification performance was validated. For preliminary results I focus on two themes: abiotic stress tolerance (e.g. cold tolerance) representing sustainability mission and aesthetics (color and smell) representing commercial interests (as a control).

To analyze the relationship between public sustainability missions and plant traits, I use two binomial logistic regressions, one for each plant trait dependent variable (abiotic stress resistance and aesthetics). These models test whether the probability of observing a specific trait within a PP abstract is predicted by the innovator type (public or private), different time periods, and their interaction. The time periods are offset by ten years from the year designated as the mission transition (1991) to account for the average time it takes plants to be bred and patents to be issued. In these regression models the intercepts are set to vary based on crop type (n = 8) to account for differences for crops' different needs and regional associations.

Results: Preliminary results suggest that public innovators prioritize different traits than private innovators, and in favor of sustainability-related traits, but not in the time periods expected. Between 1930-2000, before the sustainability mission was mainstreamed into the USDA, public innovators were slightly more likely than private innovators to breed for abiotic stress tolerance (sustainability trait). In the same period, public innovators were much less likely to breed for aesthetics (consumer preference traits). Then between 2001-2022, when we'd expect public innovators to focus on abiotic stress tolerance traits in service of the sustainability mission, the model finds that private innovators focus on abiotic stress tolerance equally as much. Moreover, the probability that public innovators' focus on these traits has decreased compared to the previous time period. Public innovators' attention to consumer preference traits like aesthetics, however, increased in the 21st century.

In summary, instead of seeing public innovators respond to the sustainability mission shift of the 21st century, we see a decline in abiotic stress resistance breeding and an increase in consumer preference traits like aesthetics. If we take results at face value, it appears that public innovators are not innovating in line with their missions, perhaps due to the increasingly enterprising university model.

ABSTRACT. Background and rationale. The pressing need to pollute less (and the urgency to mitigate negative effects stemming from climate change) has sparked widespread attention about the set of green oriented innovation strategies currently being implemented by private companies. Green innovation is generally defined as those environmentally friendly actions that tend to have an impact over firms’ scientific projects (inventive outcome and level of product newness). These include strategies for lower consumption of given materials, initiatives to decrease energy and water use, projects for environmental damage reduction, health and safety improvement, and so forth.

While technological, market and regulatory factors have been regarded as key eco-innovation drivers, few studies have explored the extent through which such sustainable schemes can also be influenced by other relevant elements such as firms’ willingness to engage in social innovation practices. Social innovation embodies the development of new initiatives, products, processes or services that are suitable to address ongoing societal challenges. Over recent years, these practices have been highlighted as novel aspects that might also shape innovative business strategies. They relate to factors like social entrepreneurship, projects conducted by non-profit organizations along with initiatives for corporate social responsibility. Against this backdrop, our research aims to explore the role played by social innovation in also shaping sustainable planning through a firm-level approach.

Data and Methods Chilean innovation surveys biannually divulged by the country’s statistical office (Instituto Nacional de Estadísticas, INE) constitute our main source of micro-level information. Such datasets report many innovation related variables including predominant type of innovation strategy being pursued (product, process, organizational and marketing), composition of R&D expenditures, availability of skilled workers, policy instruments to support scientific activity, among others. They also report variables accounting for the relevance to pursue eco-innovation strategies as well as future intentions to engage on social innovation. Our research will rely on a recent pooled sample of firms (generated by INE) which contains standardized micro data from the 11th and 12th wave of the country’s innovation datasets that comprehend the years between 2017-2018 and 2019-2020, respectively.

To adequately meet our empirical objectives, this research implemented an ordinal logit regression analysis. This econometric procedure entails examining the resulting relationship between a categorical (ordered variable) and a prearranged set of independent predictors. This is our case since Chilean companies perception about the relevance of sustainable activities are reported within our dataset as ordinal variables (with 1 being irrelevant and 4 highly important). Following prior empirical studies on Chile, we consider to two type of eco-innovation strategies as dependent variables: 1) resource efficiency (which entails plans to reduce energy consumption and the use of given materials), and 2) sustainable sensitiveness (which pertain to projects for environmental damage reduction as well as for health and safety improvement). To have additional insights into the importance of green initiatives of any sort, our research also introduced a third dependent variable (full eco-innovation) which is herein operationalized as the sum of the two recently advanced indicators.

Social innovation practices are operationalized as a binary independent variable underscoring firms’ willingness to embark on such type of strategies in the future. Factors such innovation capabilities (sum of inventive strategies followed by the organization), reliance on cooperative agreements, type of ownership (foreign or domestic), production for exports, access to government aid, sector of economic activity, prior experience on eco-innovation strategies, together with indicators for firm size and age are too included as additional predictors in our ordinal regression. It is also worth noting that (to account for potential endogeneity concerns) our dependent variables are solely examined in time t (2020), while each of the previously independent regressors are correspondingly evaluated at time t-1 (2018).

Econometric results. Our general regression results can be summarized as follows: 1) Future intentions to conduct social innovation practices (reported in period t-1) positively shape the deployment of full eco-innovation schemes (executed in time t). Nevertheless, when each green oriented projects is individually addressed, we note that social innovation only seems to induce higher sustainable sensitiveness, thereby standing as irrelevant for the deployment of resource efficiency strategies. 2) Higher innovation capabilities, increasing exporting activity and foreign ownership (MNEs) stand as powerful predictors of either type of eco-innovation activity herein considered. 3) Participation in cooperative agreements with other research institutions generates a stronger impact on resource efficiency projects than on those that are related to sustainability. 4) Access to government aid increases firms’ positive perception on sustainable sensitiveness practices to a greater extent than they do for resource efficiency. 5) Prior knowledge on resource efficiency projects (as reported by the firm in time t-1) influences the execution of either individual green innovation activity in period t. Nevertheless, former experience on sustainable initiatives only motivates the pursuit of similar environmental strategies of this sort (in time t), and reports no relevant effect on resource efficiency schemes.

Conclusions and points for discussion. These results have strong policy implications as they highlight the fact that green innovation activities at developing countries might be differently affected by technology, market and regulatory elements depending on the environmental priority that firms seek to improve on. While stressing the idea that social innovation might raise firms’ awareness to also pursue strategies for sustainable sensitiveness (i.e., environmental damage reduction), we also observe that such type of societal-related initiatives do not necessarily spark interest in pursuing other more complex green oriented projects .The fact that energy efficiency projects were found to heavily rely on firms’ prior experience on these type of initiatives (as well as on the presence of cooperative agreements) underscores that search for lower costs stands as a more relevant determining factor than societal awareness over this specific endeavor.

ABSTRACT. The transition to a sustainable energy future necessitates large-scale energy transformation. The transition must not only achieve climate targets without unnecessary disruption to economic activity, but must also be socially and politically acceptable. Energy adaptations that are perceived to impose an undue burden on consumers will not be implementable. Various interventions to increase policy acceptability have been conducted, but the most effective communication channels and the duration of their effects remain uncertain. This work deciphers the role of information provision on the public acceptability of key energy and climate policies and the persistence of these effects over time. We aim to answer three questions:

• Do people change their pre-existing acceptance of key energy and climate policies when presented with new information regarding the policies’ outcomes? • What type of information (environmental, economic or equity outcomes) is most likely to influence change in acceptability? • Are information provision effects sustained over time?

To answer these questions, we conduct an innovative behavioral experiment featuring a unique feedback mechanism based on real-time information provided by a macroeconomic model. We conduct a follow-up study after six months, and report the conditions under which information provision effects are persistent.

Our case study is Ireland, a European country with ambitious climate policies, such as the Climate Action Plan, which considers increased deployment of sustainable energy and energy efficiency technology, together with increases in carbon taxation. With a nationally representative sample of 1,000 adults, we run the behavioral experiment and ask participants to state their preferences and acceptance of five policies that are included in the 2021 Climate Action Plan. These are:

1. Increasing renewable electricity up to 80% by 2030. 2. Increasing carbon taxes to 100 Euros (approx. $105/ton) by 2030- compared to the previous rate of 33 Euros ($35)/ton. 3. Distributing carbon tax revenues to households, in the forms of social protection payments and assistance for energy efficiency improvements. 4. Increasing the number of electric vehicles to 1,000,000 by 2030. 5. Increasing the number of heat pump installations to 600,000 by 2030.

Immediately afterwards, each participant receives tailor-made information about the long-term environmental, economic and equity outcomes of the policies they have accepted. The latter outcomes are predicted by the Ireland Environment, Energy and Economy model, our established computable general equilibrium model that simulates the Irish economy in its entity. We model 32 possible scenarios, with each scenario representing the implementation of some- or all- of the policies. Depending on their answers, participants immediately see on the screen the outcomes of a unique scenario in the form of infographics. If participants originally indicated that they agree (disagree) to some extent with a specific policy being implemented, the scenario will simulate that the policy is (is not) implemented. Furthermore, if they originally indicated that the majority (minority) of carbon tax venues should be distributed to support social protection or climate targets, the generated scenario will reflect that too. After viewing the results, the participants have the choice to update their preferences if they wish to, up to two times, and see how their updated answers affect expected outcomes.

Key information about policy outcomes include: • Environmental outcomes: changes in carbon dioxide emissions. • Economic outcomes: changes in unemployment and in cost of living. • Equity outcomes: changes in real disposal income for low-income and for high-income households.

Through transition matrices, we assess the probabilities of participants to assign higher scores after the intervention, compared to those they assigned before the intervention. To explore the relative effect of each information type, participants are randomized so that 25% of them receive information about the environmental outcomes only, 25% receive information about the economic outcomes only, 25% about the equity outcomes only, and other 25% about all three outcomes. Other interests include whether the existing or updated levels of policy acceptance depend on: prior knowledge about these policies, perceived effectiveness of the techno-economic innovations the policies entail, perceived importance of each outcome, perceived behavioral control of each outcome, and most trusted source of information about climate change. We employ panel regression models to assess the role of each parameter in shaping acceptance of the policies. We also study the heterogeneity of effects across various socio-demographic groups. After six months, we conduct a follow-up study to explore whether participant preferences, as given at the end of the behavioral experiment, have been modified, or not, and to what extent.

Both our transition matrices and regression models indicate that informing the public about the impacts of all climate policies significantly increases their acceptance. It is noticeable that even participants with very low initial levels of acceptance display noteworthy probabilities for high levels of acceptance post-intervention. Furthermore, we found that changes in policy acceptance are associated with a combination of information provision, socio-demographics and personal values. Providing combined information about three types of outcomes (environmental, economic, and equity) is often more effective than information about a single type. Among the various types of information, environmental information is found to be the most influential, followed by information about equity outcomes. Perceived behavioral control of the environmental outcome was another significant predictor that emerged. After six months, we observe that intervention effects have been sustained mostly for participants who had demonstrated mid-level acceptance of the policies at the end of the initial experiment. On the other hand, participants who had demonstrated high levels of acceptance at the end of the initial experiment, now demonstrate lower acceptance.

Overall, this work provides valuable insights into designing and communicating climate policies that are likely to gain long-term social acceptance. Our results create optimism that public acceptance of policies for energy and climate can be increased through carefully-designed and well-communicated information provision. This is particularly noteworthy in the case of acceptance of carbon taxes, considering that the latter are traditionally considered unpopular. At the same time, our results demonstrate the difficulty of sustaining high levels of acceptance over time without re-enforcement, and highlight the need for continuing efforts to maintain acceptance through information provision strategies.

ABSTRACT. Abstract: Despite substantial policy efforts aimed at developing regional innovation systems (RIS), our understanding of institutional factors that promote synergy and integration at the regional scale is limited. To address this gap, we analyzed historical patterns of research co- production within and across California (CA) and Texas (TX), two RIS that together account for >5% of global scientific production. This predominance is largely attributed to the University of California and the University of Texas, two multi-campus university systems (MUS) that feature distinct configurations of institutional specialization. We exploit these differences to analyze four institutional assortativity channels that foster system-level synergies: institutional proximity, prestige, homophily, and specialization. Our results show that regional integration is mediated by the alignment of institutional specialization and moderated by institutional homophily, which thereby highlights MUS as valuable sources of institutional redundancy and variation that generate a large combinatorial space of potential multi-university research synergies. This configurational advantage is supported by results showing that research co- produced with a premier regional university and an international university feature a 21% (CA) and 25% (TX) citation premium. Hence, RIS administrators can extend productive capacity and global impact by identifying institutional synergies that optimize for both regional integrity and international competitiveness.

Summary: Regional innovation systems (RIS) are the foci of substantial science policy initiatives aimed at strategically integrating national and pan-national innovation systems by supporting the co-production of knowledge and the circulation of high- skilled labor. Chief among the institutional and geographic fixtures that define regional innovation systems (RIS) are research universities (RU) [2–4], which represent sources, anchors and hubs [5] across vast multi-level networks of human, social and intellectual capital. Here we present a framework for understanding how institutional assortativity – broadly defined as institutional similarities that support research co- production affinities that persist over distance and time – promotes RU ecosystem integration at the regional scale. Accordingly, in what follows we analyze four distinct channels of institutional assortativity: proximity, prestige, homophily, and specialization.

Regarding regional proximity, we focus upon California and Texas, two geographically and politically distinct regions within the US innovation system. A principal component of our RU sample are the 10 institutions belonging to the University of California (UC) system and the 12 institutions belonging to the University of Texas (UT) system. We complement these two public multi-campus university systems (MUS) with six prominent private universities that form a non-MUS comparison group. Together, the 28 RU in our sample are affiliated with roughly 3 million publications collected from Clarivate Analytics Web of Science Core Collection (WOS), representing >5% of publications indexed by WOS over the sample period 1970-2020.

Regarding institutional prestige, previous research emphasizes premier universities because they generally produce research in large volumes [6, 7]. Instead, our focus at the regional scale exposes the broad distribution of institutional sizes, including the more numerous yet less prolific public universities that nevertheless generate commensurate research and education contributions on the aggregate [1]. There is considerable variation in research production among UC and UT institutions, which contributes to broader prestige hierarchies observed in science [7–9.

Another key characteristic of MUS is the replication of organizational principles and policies across various locations, meaning that certain aspects of institutional environment can be considered approximately fixed across the different campuses (e.g. tenure and promotion policy). Such organizational commonalities also promote homophily, whereby social groups tend to initiate and persist via the identification of shared attributes, values and experiences. The great number of graduates generated by MUS thereby generates several distinct avenues that reinforce institutional homophily, including but not limited to collaboration, employment, student admissions and brand equity. Thus, the direct and indirect network effects generated within and across MUS campuses fosters a considerable organizational advantage relating to a fundamental objective of RU administration, which is to attract and retain prolific scholars [3].

And finally, we exploit the different organizational configurations of the UC and UT systems to understand how institutional specialization conditions the combinatorics of institutional alignment. Notably, the UT is comprised of a 7 campuses representing traditional multi- disciplinary RU, coupled with 5 campuses that are specialized biomedical and health science centers. Conversely, 9 of 10 UC campuses are traditional RU, the exception being UCSF which specializes in biomedical and health science research and graduate-level educational programs, akin to UT health science centers. From an organizational standpoint, the bureaucratic complexity of managing diversified universities [10] is likely to compound in MUS, especially if resource allocation and investments across the system are guided by principles of competitiveness, equity and transparency. The varying levels of institutional specialization among MUS members create trade-offs between redundancy and diversity, especially concerning the effectiveness of a uniform governance approach.

Our study thus targets a longstanding research question owed to Aristotle and tailored to modern RU ecosystems: How does variation among institutions manifest in operational synergies that enhance the productive capacity of the innovation system as a whole? Our findings emphasize the importance of public investment in multi-campus university systems, which serve as foundational pillars within RU ecosystems. From a stakeholder perspective, our framework may be instructive to institutional research development offices tasked with identifying strategic investments and multi-university funding opportunities.

References:

[1] Madhavan G, et al. (2020) NAE: The Bridge on Complex Unifiable Systems (National Academies Press). [2] National Research Council (2012) Research universities and the future of America. [3] Rouse WB (2016) Universities as complex enterprises (John Wiley & Sons). [4] Rouse W. B., et al. (2018) Modeling research universities. PNAS. [5] Owen-Smith J (2018) Research universities and the public good. (Stanford University Press). [6] Adams JD, et al. (2005) Scientific teams and institutional collaborations: Evidence from US universities, 1981–1999. Research policy. [7] Jones BF, et al. (2008) Multi-university research teams: Shifting impact, geography, and stratification in science. Science. [8] Clauset A, et al. (2015) Systematic inequality and hierarchy in faculty hiring networks. Science Advances. [9] Wapman KH, et al. (2022) Quantifying hierarchy and dynamics in US faculty hiring and retention. Nature. [10] Walsh JP, Lee YN (2015) The bureaucratization of science. Research Policy.

ABSTRACT. Introduction The transition to a circular economy (CE) has gained momentum as a response to pressing sustainability challenges and in particular the need to decouple economic growth from resource consumption. Circular start-ups represent a unique category of firms within this transition, as they build circular business models from the scratch, focusing on sustainability-driven innovation. Unlike incumbents that may incrementally adopt circular principles, these start-ups embed circularity at the core of their business model, aiming to minimize waste and maximize resource efficiency. They are thus pivotal actors in advancing circular economy goals, addressing both environmental and socio-economic challenges. This study systematically reviews the literature on circular start-ups, offering a consolidated understanding of their role, challenges, and contributions within circular ecosystems. In doing so, we develop a framework to guide future research and policy agendas, with relevance for science, technology, and innovation (STI) policies.

Research Questions Despite growing attention in the literature, the unique role of circular start-ups in driving the transition to a circular economy (CE) remains underexplored. This study seeks to fill this gap by addressing two key research questions. First, it examines the specific roles circular start-ups play within CE ecosystems, with a particular focus on their contributions to value co-creation and collaborations. Second, it investigates the challenges these start-ups face due to their newness and size, exploring how these factors influence their ability to contribute effectively to CE objectives.

Methodology This study utilizes a systematic literature review to investigate the role of circular start-ups in circular economy (CE) transitions. Using the Scopus and Web of Science databases, the review focuses on publications from the past decade to capture recent developments and emerging research trends. The analysis follows a two-stage process. First, a thematic analysis was conducted to categorize findings related to the roles, definitions, and ecosystem interactions of circular start-ups. Second, a lexical analysis was performed to uncover underlying structures, keywords, and research gaps across the 63 selected articles. This dual approach not only identified dominant research themes and emerging subtopics within the literature but also highlighted under-researched areas of significance to both scholars and policymakers.

Findings and Contributions Our analysis makes three key contributions to advancing the discourse on circular start-ups and their role in the transition to a circular economy (CE). First, we offer definitional clarity by refining the concept of circular start-ups and distinguishing them from other business models within the CE domain. Circular start-ups are defined as small, innovation-driven enterprises that prioritize circularity principles at the core of their business models, focusing on value creation through resource recovery, closed-loop systems, and collaborative value chains. This definition highlights their distinct contributions and provides a robust foundation for future empirical studies. Building on this, we propose a conceptual framework positioning circular start-ups within the broader CE transition. The framework identifies their core functions, such as resource optimization, product life extension, and the development of closed-loop systems. It also explores their ecosystem roles, including interactions with large corporations, government bodies, and research institutions to foster knowledge transfer and co-create solutions. Additionally, it examines the critical interfaces where policy interventions can address structural barriers and stimulate growth, particularly in areas like access to funding and regulatory support.

Second, we analyze the roles circular start-ups play in value co-creation and ecosystem integration. These start-ups serve as vital bridges between knowledge and practice within innovation ecosystems. Through partnerships and networks, they engage in collaborative innovation, co-creating value with larger corporations, research institutions, and government agencies. Despite their potential, circular start-ups face significant challenges, including limited resources, lack of economies of scale, and regulatory uncertainty, which can hinder their ability to drive systemic change independently. By fostering cross-sector partnerships, they can amplify their impact, influencing supply chains and contributing to sector-wide transitions. We argue that policies encouraging collaborative models and ecosystem building—particularly those that facilitate risk-sharing across public and private entities—can help these start-ups overcome barriers and create a conducive environment for experimentation and innovation.

Finally, our study concludes with a research agenda and policy recommendations that emphasize the need to support circular start-ups in scaling their impact. Future research should explore alternative funding mechanisms, such as green bonds and impact investment, to address financial barriers. It should also focus on designing and evaluating policy instruments tailored to circular start-ups, including grants, subsidies, and public procurement policies. Additionally, enhancing frameworks to measure the multi-dimensional impacts of circular start-ups, including their environmental, economic, and social contributions, is crucial. We call for interdisciplinary research that integrates insights from sustainability science, entrepreneurship, and innovation studies, aligning with broader goals of fostering sustainable development.

This study is highly relevant to the 2025 Atlanta Conference on Science and Innovation Policy, as it directly engages with the theme of addressing complex global challenges. Circular start-ups represent a form of science- and technology-driven entrepreneurship capable of advancing sustainability goals across diverse contexts. By exploring key STI policy dimensions, such as regulatory frameworks, financing, and evaluation, this research aligns with the conference's focus on shaping policies that support inclusive and sustainable innovation. Furthermore, the emphasis on ecosystem integration and collaborative models underscores how circular start-ups exemplify impactful and inclusive innovation, contributing to economic resilience and environmental stewardship. Our findings and proposed agenda offer actionable insights for researchers and policymakers to create enabling conditions for these start-ups, advancing the transition to a circular economy and supporting broader sustainability transitions.

Conclusion Circular start-ups are crucial yet under-researched actors within the CE transition, facing unique opportunities and challenges. By offering a comprehensive review of current knowledge and mapping a future research agenda, this study provides insights to guide policy and academic discourse. Understanding and supporting circular start-ups can strengthen their role in innovation ecosystems, contributing to sustainable development goals and addressing global challenges. This expanded understanding is essential for STI policies that foster sustainable, inclusive, and resilient economies—a key aim of the 2025 Atlanta Conference on Science and Innovation Policy.

ABSTRACT. Abstract: Because the distribution of knowledge is uneven, multi-location firms build or acquire establishments in various locations in order to gain access to the local knowledge base (Beugelsdijk & Mudambi, 2013; Kafouros et al., 2018). In this way, the multi-location firm can access localized pools of knowledge that deepen its existing technology base or provide a means to diversify into high level technologies (Alcácer & Zhao, 2012; Frigon & Rigby, 2024). Less attention has been given to the impact of multi-location firms on local knowledge development. This study explores the effects of new multi-location firms’ establishments on local knowledge development through the lens of the dynamic regional innovation system. Specifically, we examine: (1) whether multi-location firms foster knowledge development by local firms and universities, and (2) how internal linkages within multi-location firms might influence local knowledge development. We focus on the semiconductor industry as our empirical setting, given its reliance on technological advancement for success (Wang & Zhao, 2018). To identify relevant companies, we select those that self-identify as operating within the semiconductor industry based on their membership in the Semiconductor Industry Association (SIA), following the approach of Feldman and Lendel (2010). Our investigation spans the innovation activities of 28 companies from 2016 to 2020, analyzing their patents, geographic locations, and establishments to gain insights into their contributions to technological progress. The data for this study comes from two main sources. First, we use patents from the USPTO and academic articles from Web of Science to analyze the knowledge flow and collaboration among local industries, universities and multi-location firms, as well as the internal linkages within multi-location firms. Second, we incorporate data from Crunchbase and OpenCorporates to track the subsidiaries and hierarchy of multi-location firms. By integrating these datasets, we match patents and articles to both inventors’ and firms’ R&D locations, allowing us to map the geographical and organizational structure of knowledge flows among local industries, universities and multi-location firms through co-authorship and citations. To measure local knowledge development, we focus on three key indicators: knowledge diversity, specialization, and complexity. These indicators are assessed using patents and papers from the perspectives of both local universities and industries, as both serve as critical drivers of innovative activity (Feldman & Lendel, 2010). This study contributes to a deeper understanding of how multi-location firms interact with and shape regional innovation systems. It uncovers the interplay between organizational structures and regional innovation systems, offering valuable insights into the dynamics of knowledge sharing in the context of multi-location operations. References Alcácer, J., & Zhao, M. (2012). Local R&D strategies and multilocation firms: The role of internal linkages. Management Science, 58(4), 734-753. Beugelsdijk, S., & Mudambi, R. (2013). MNEs as border-crossing multi-location enterprises: The role of discontinuities in geographic space. Journal of International Business Studies, 44, 413-426. Feldman, M. P., & Lendel, I. (2010). Under the lens: the geography of optical science as an emerging industry. Economic geography, 86(2), 147-171. Frigon, A., & Rigby, D. L. (2022). Where do capabilities reside? Analysis of related technological diversification in multi-locational firms. Regional Studies, 56(12), 2045-2057. Frigon, A., & Rigby, D. L. (2024). Geographies of Knowledge Sourcing and the Complexity of Knowledge in Multilocational Firms. Economic Geography, 100(4), 329-350. Kafouros, M., Wang, C., Mavroudi, E., Hong, J., & Katsikeas, C. S. (2018). Geographic dispersion and co-location in global R&D portfolios: Consequences for firm performance. Research Policy, 47(7), 1243-1255. Lavoratori, K., Mariotti, S., & Piscitello, L. (2020). The role of geographical and temporary proximity in MNEs’ location and intra-firm co-location choices. Regional Studies, 54(10), 1442-1456. Wang, S., & Zhao, M. (2018). A tale of two distances: a study of technological distance, geographic distance and multilocation firms. Journal of economic geography, 18(5), 1091-1120.

ABSTRACT. Nascent industries compete with incumbent industries. In the process of this competition, conflicts may intensify between the nascent and incumbent industries. Recently, the emergence of digital platform technology has amplified this competition. This is because nascent industries based on digital platforms are encroaching on existing markets at a faster rate than ever before. As competition intensifies, it sometimes becomes a social issue. A typical example of this is the mobility service industry. Nascent industry participants launch products and services with new functions that an incumbent industry does not provide. Nascent industries offer customer oriented products and services as a differentiation strategy. These new products and services seem to spread rapidly, but there are still many difficulties for nascent industries to overcome incumbent industries. Incumbent industries have long created industrial and technological systems and infrastructures to suit their respective business models. Consequently, incumbent industry participants have strong market power and can impede the diffusion of new technologies and business models (Adner & Kapoor, 2016; Chang & Wu, 2014; Geels, 2004). The legitimation process plays a key role in the growth of a nascent industry. Legitimation refers to the process by which nascent industry products and services are widely accepted and recognized as legitimate by society (Aldrich & Fiol, 1994). Specifically, it is necessary for the participants in the nascent industry to form their own identity through collective cooperation to be recognized as a visible entity in society (Georgallis et al., 2019; Liu & Guenther, 2024). Furthermore, they can undermine incumbent industry’s legitimacy by giving a negative image to the incumbent industry, thereby creating a de-legitimation effect and gaining public support (Ferns et al., 2022; Thomas & Ritala, 2022). Nascent industry participants utilize a variety of strategies to legitimize a nascent industry; however, their strategies may not be effective in some cases. The effectiveness of a legitimation strategy can vary depending on social and economic conditions of competing incumbent industries. However, prior studies mostly have examined the legitimation strategy against socio-economically advantaged incumbents, referring to a group with strong social status and economic resources. For example, Georgallis et al. (2019) examined that the solar industry, as a nascent industry, was able to significantly increase its legitimacy by competing with a socio-economically advantaged incumbent (i.e., existing fossil fuel energy firms), consequently gaining government support. The incumbent fossil fuel energy industry has strong market and political position so that the government can support the solar energy industry without social criticism. Another example is the case of the finance industry. Prior studies also analyzed how counter parts gain social supports by criticizing socio-economically advantaged incumbents in the finance industry (Budd et al., 2019; Roulet, 2015). However, little research has been conducted on the case of socio-economically disadvantaged incumbents (i.e., a group with weak social status and economic resources). This study aims to analyze the legitimation strategy adopted by nascent industry participants and the counter-strategy of incumbent industry participants to investigate the effects of social and economic conditions of incumbents. In particular, we examine collective cooperation and de-legitimation of nascent industry participants against socio-economically disadvantaged incumbent competitor. For this purpose, we use the Korean mobility service industry as a case study. In recent years, the Korean mobility service industry has experienced the exit of various carpooling services. A prominent ride-sharing service, Tada Basic, also exited the market due to the competition with the incumbent taxi industry. In particular, Tada Basic’s exit became an important social issue in Korea. In the mobility service domain, we set i) the taxi industry as the incumbent industry, and ii) mobility platform industries that provide carpooling and ride-sharing services as the nascent industry. We conducted a longitudinal case study to investigate the events that occurred among diverse stakeholders in the mobility service industry from August 2013 to June 2023 (Yin, 2009). A longitudinal case study based on a historical point in time was deemed appropriate due to the lack of research on the competitive relationships between early players in the industry and new entrants, and the lack of quantitative data as an industry is just beginning to take shape. Our results show that nascent industries (i.e., ride-sharing and carpooling industries) invest in their legitimation process to increase the social acceptance of their new services while they adopt a differentiation strategy by introducing various product innovations to provide their users with new benefits that incumbent industries have not provided. To facilitate this legitimation process, nascent industry participants attempt to develop collective cooperation and undermine the legitimation of the incumbent taxi industry (i.e., de-legitimation) (Ferns et al., 2022; Georgallis et al., 2019; Liu & Guenther, 2024). However, their strategies were not effective against socio-economically disadvantaged incumbents. The current study makes two contributions. First, we analyze the process of collective cooperation when nascent industry participants compete with socio-economically disadvantage incumbent. Previous studies highlighted that nascent industry participants share a common goal of overcoming incumbents, and therefore, it should be relatively easy for them to establish collective cooperation against incumbents (York et al., 2016). However, nascent industry participants’ diverse interests and opinions cannot be converged when they compete with socio-economically disadvantaged incumbents. In this situation, nascent industry participants should adopt a long-term perspective from the beginning of the industry formation and undertake continuous efforts to coordinate their interests (Choi et al., 2011). Second, we analyze the instances in which a de-legitimation strategy works on socio-economically disadvantaged incumbents. Many prior studies demonstrated a merit of nascent industry participants when deploying their de-legitimation strategies on socio-economically advantaged incumbents. However, when incumbent participants are a socio-economically disadvantaged group, for example, the Korean taxi industry, de-legitimation strategies work as a complex social process. In this case, the effects of de-legitimation strategies differ from those observed in the conventional cases analyzed in previous literature. This study shows that nascent industry participants must analyze the social and economic conditions of incumbent industry participants when implementing de-legitimation strategies against incumbent industry participants.

ABSTRACT. This paper examines the relationship between occupational norms and different forms of research misconduct. In recent years, once regarded as paragons of collegial organization (Blau, 1994), universities have increasingly adopted bureaucratic structures (Lee & Walsh, 2021), leading to a more metrified and commercialized system of knowledge production (Berman, 2011). At the same time, the scientific community has become more demographically diverse, raising the question of whether a single normative framework can adequately account for researchers’ varied motivations and behaviors (Woo & Walsh, 2024). In this paper, we provide evidence suggesting that different research environments, shaped by either Mertonian norms or bureaucratic environments, predict distinct forms of misconduct. Under Mertonian norms, researchers internalize the values of sciences in pushing the boundaries of knowledge to gain social recognition rather than merely fulfilling institutional metrics (Merton, 1973). In contrast, researchers operating in a bureaucratized science are more likely to publish for the sake of accumulating publications, which directly count toward their performance evaluations (Biagioli & Csiszar, 2020). Such an environment may foster a high likelihood of goal displacement, turning publication into a means to an end. More specifically, we posit that research misconducts committed by researchers from institutions or countries more aligned with the Mertonian norms are more likely to engage in “bold” research misconduct, deliberately altering evidence to publish in top journals (Park, 2020; Reich, 2009). This tendency paradoxically may stem from their strong adherence to Mertonian norms, particularly their emphasis on novelty (Zuckerman, 1988). Meanwhile, researchers from more bureaucratized research settings, who are not traditionally enculturated in the core, are less likely to share Mertonian norms and are more susceptible to institutional pressures to maximize metrified performance, such as the number of publications. We posit that this group is more prone to “petty” research misconduct, which may not critically damage core knowledge. Examples include plagiarism, fake peer review, and the use of paper mills. We test our hypotheses using a series of regressions based on over 50 thousand retracted articles from RetractionWatch and bibliographic information obtained from OpenAlex. First, retracted papers from scientifically established countries, institutions, and prestigious journals, which we assume to be associated with traditional Mertonian norms, are more likely to involve bold misconduct, such as fabricating data, materials, or results. Meanwhile, retracted papers from less developed countries, institutions, and less prestigious journals are more likely to be associated with petty misconduct, such as plagiarism, paper mills, and peer review manipulation. Our findings demonstrate that research environments shaped by Mertonian norms versus bureaucratic pressures produce distinct forms of misconduct. Researchers often engage in “bold” misconduct in Mertonian settings aimed at high-impact recognition. In contrast, those in bureaucratized contexts are more likely to commit “petty” offenses driven by institutional performance metrics. These insights carry important policy implications for universities and funding bodies, which should tailor their oversight and evaluation criteria to scientists’ unique pressures. We provide detailed policy recommendations and also address emerging concerns over the misuse of large language models (LLMs) in paper production, particularly in understanding which researchers, based on their normative environment, are more likely to use LLMs for misconduct.

References: Berman, E. P. (2011). Creating the market university: How academic science became an economic engine: Princeton University Press. Biagioli, M., & Csiszar, A. (2020). Gaming the Metrics: Misconduct and Manipulation in Academic Research: Mit Press. Blau, P. M. (1994). The organization of academic work: Transaction Publishers. Lee, Y.-N., & Walsh, J. P. (2021). Rethinking Science as a Vocation: One Hundred Years of Bureaucratization of Academic Science. Science, technology, & human values, 0(0), 01622439211026020. doi:10.1177/01622439211026020 Merton, R. K. (1973). The normative structure of science. In N. W. Storer (Ed.), The sociology of science: Theoretical and empirical investigations (pp. 267-278): The University of Chicago Press. Park, B. S. (2020). Making matters of fraud: Sociomaterial technology in the case of Hwang and Schatten. History of Science, 58(4), 393-416. Reich, E. S. (2009). Plastic fantastic: How the biggest fraud in physics shook the scientific world (Vol. 1): Macmillan. Woo, S., & Walsh, J. P. (2024). On the shoulders of fallen giants: What do references to retracted research tell us about citation behaviors? Quantitative Science Studies, 1-30. doi:10.1162/qss_a_00303 Zuckerman, H. (1988). The sociology of science. In Handbook of sociology. (pp. 511-574). Thousand Oaks, CA, US: Sage Publications, Inc.

ABSTRACT. Often scientists (here including physical scientists, engineers, and social scientists) are at a loss to understand why ordinary citizens, including many who have extensive formal education, reject the data, findings, and theories produced by scientists, even when scientists view these forms of knowledge to be uncontroversial and consensually validated. The question of why non-scientist repudiate science is examined here, in depth, drawing from history, data, and current social, political and technological factors. My objectives include: (1) to provide a better framing by distinguishing between lack of trust (general) in science (general) and repudiation of science (specific), (2) to develop ideas about the circumstances in which science is and is not "a superior way of knowing" compared to other forms of knowledge (e.g., personal knowledge, craft knowledge, folk knowledge, moral knowledge) and why, (3)) to develop propositions about factors causing repudiation, including judgments as to criteria for assessing whether such repudiation is or is not "justified", (4) to provide a preliminary integrative theory about repudiation of science, and, finally, (6) to consider the implications of repudiation for public policy and some steps that scientist and policymakers may consider as they seek to mitigate socially harmful repudiation of science.

ABSTRACT. The language used to describe political ideologies varies significantly across nations, reflecting diverse meanings and interpretations. This section explores the spectrum of political ideology groups. Here, the term “conservative” encompasses a blend of ideological stances: a push for minimal government intervention in markets (neoliberalism), adherence to traditional moral values—particularly regarding sexual and marital norms—and policies rooted in nationalism or nativism. Many nations feature both moderate conservative factions and far-right parties, or these perspectives coexist as distinct wings within a broader conservative movement. While moderate conservatives often demonstrate greater openness to energy transition initiatives and exhibit varied views on clean energy, far-right groups frequently dismiss climate science, advocating instead for the continued reliance on coal, fossil fuels, and natural gas (Hess and Renner 2019; Mayer 2019).

This paper discusses the spectrum of clean energy conservatism by analyzing people’s attitudes towards renewable energy (RE) and nuclear energy (NE). Clean energy conservatism has already been discussed in the literature. For example, Lee and Hess (2024) have discussed how conservative party members are connected with fossil fuel industries through lobbying. Therefore, the energy transition, for conservative party members is tightly connected with economic incentives. Similarly, Hess and Pride Brown (2017) found that the key frames for conservative party members that support clean energy tended to focus on job creation, economic innovation, affordability, and tax revenue. Following from this logic, this paper asks the following research question: how is clean energy conservatism displayed in South Korea, and how strong are the connections between the clean energy conservatives and economic benefits?

This paper is analyzing two types of energy, namely RE and NE, and how people’s attitudes towards them tend to change depending on their political inclination. It uses survey data collected in South Korea in 2024. The respondents of the survey were asked to rate their support for RE and NE using Likert Scale. Then, they were also asked to rate to what extent they support RE and NE because of environmental reasons and/or economic reasons. This question is particularly interesting in the context of South Korea because the country has had a strong divide in terms of types of energy support depending on the administration in power – with the conservative administrations providing economic incentives and policy support to NE and suppressing RE, whereas the liberal administrations tended to take the opposite route. Therefore, how clean energy conservatism narratives are played out in South Korea is an interesting dimension that adds to the existing literature.

The preliminary findings indicate that the conservatively identifying individuals tend to prefer nuclear energy compared to renewable energy. Additionally, it found that conservatively identifying individuals supported NE because they believed it to be more beneficial to the Korean national economy and more important for the clean energy transition compared to RE. The findings challenge the prevailing narrative that clean energy conservatism is primarily tied to economic incentives, highlighting instead the complex role of environmental conditions in shaping conservative support for NE.

This study contributes to the existing literature on clean energy conservatism by offering a nuanced exploration of how political ideologies shape attitudes towards RE and NE, with a specific focus on South Korea. Although prior research has established the economic frames that conservatives use to support clean energy, this study broadens the discussion by examining the interplay between economic and environmental justifications for energy preferences.

The findings of this study have significant policy implications for advancing clean energy transitions in politically polarized contexts. Policymakers should recognize the distinct frames through which conservative groups approach energy policy, particularly their economic and environmental motivations for supporting nuclear energy over renewable energy. In South Korea, where political ideologies heavily influence energy policy, designing targeted strategies that align with conservative priorities—such as emphasizing the economic and environmental benefits of renewable energy—could help bridge the ideological divide. Additionally, fostering bipartisan support for energy transitions may require reframing renewable energy policies to highlight their contributions to national economic growth, energy security, and job creation, which are key concerns for conservative groups. These strategies could encourage more inclusive and sustainable energy policies, reducing the reliance on divisive narratives and promoting a unified vision for clean energy development.

ABSTRACT. Introduction The automated vehicle (AV) industry is steadily expanding its market size with attracting public attention. As the technology advances rapidly, the situation in which related legislation and institutional arrangements lag behind is problematic. While the public expects AV technologies to reduce accidents, people also feel uncertainty or anxiety, and demand democratic preparations for securing safety, equality, sustainability and privacy. Therefore, I expect that public trust in government that make these preparations is also related to the acceptance of AVs, and state the research question: How do existing institutions lead to greater trust in AVs? This paper conducts a large-scale survey on AVs and applies the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) model to measure correlations (Pavlou, 2003). By identifying the effect of government branches, this analysis preemptively explains the possibility that AV acceptance will be change politically in a democratic society, and that the AV acceptance may vary depending on an individual’s political orientation. Furthermore, it extends the current UTAUT2 model to contribute to the methodology of interpreting AV surveys.

Literature Review What the term AV refers to varies slightly depending on the context, but in this study, AVs are specifically fully autonomous and privately owned vehicles that no longer require a traditional driver. The acceptance of AV has been dealt with as an extension of the existing technology acceptance theory, so it has been closely related with the concept of trust. In general, trust can be defined as the degree to which the target's behavior is expected to be consistent with one's intention. However, Hoff and Bashir (2015) distinguish it into dispositional, situational, initial learned, and dynamic learned trust. As a result, the object of trust is changed to the concept of a system, and the elements of trust are expanded to the product, the manufacturing company that guarantees it, and the national level that manages it. For this paper, the elements that constitute public trust at the government level can be subdivided into the legislative, judicial, and executive branches (OECD, 2017; Trkman et al., 2023). According to the existing papers, it has been shown that using or extending the UTAUT2 model for AV study is reliable in revealing these new correlations (Korkmaz et al., 2022; Nordhoff et al., 2020).

Method South Korea was suitable for conducting a survey because they are not particularly quick in preparing for the era of AVs, and the division between the legislative, judicial, and executive branches are valid. The survey was conducted from October 29 to November 11, 2024, after collecting a sample of 1,864 people through a professional public opinion research institute. Taking into account the population ratio, the data also secured 575 people aged 60 or older, who are generally considered to be vulnerable to transportation system. This study assumes latent variables such as Performance Expectancy, Effort Expectancy, Social Influences, Hedonic Motivation, and Risk in UTAUT2 model for correlation analysis, which is similar to previous studies. The personal innovativeness, subjective safety, and demographic variables such as gender, age, annual driving distance, and political orientation were included in the questionnaire (de Graaf et al., 2019). Most importantly, variables of trust in companies and government branches were added, with the intention of confirming a multi-layered trust structure.

Findings Structural Equation Modeling (SEM) was used for the analysis. Primary findings support previous studies that Performance Expectancy and Social Influences most clearly affect the behavioral intention. Moreover, in addition to trust in AV, trust in manufacturing company and government was significant. However, personal political orientation was an important variable as it affects overall trust in government branches. The technology-oriented attitude called personal innovativeness was also related to behavioral intention.

Discussion and Conclusion The successful extension of the concept of trust to the company and government levels aligns with Hoff and Bashir's systemic trust perspective. On the other hand, it is noteworthy that acceptance fluctuates with the political orientations of the ruling party or individuals. In the case of Korea, the two parties have similar technology-oriented attitudes, so this issue has not yet been politicized. However, this preliminary result reveals the possibility that the ruling party's direction toward AV policy may be treated as a political issue. Conversely, if this does not become a political issue, still there is a risk that autonomous driving technology will be introduced without gaining sufficient trust. In particular, given the demand for a fair system in the judiciary, the legal debate over AVs may undermine the multi-layered trust structure. As the introduction and public acceptance of AVs involve these political tensions, this paper calls for an expanded approach in future research.

References de Graaf, M. M. A., Ben Allouch, S., & van Dijk, J. A. G. M. (2019). Why Would I Use This in My Home? A Model of Domestic Social Robot Acceptance. Human–Computer Interaction, 34(2), 115–173. https://doi.org/10.1080/07370024.2017.1312406 Hoff, K. A., & Bashir, M. (2015). Trust in Automation: Integrating Empirical Evidence on Factors That Influence Trust. Human Factors, 57(3), 407–434. https://doi.org/10.1177/0018720814547570 Korkmaz, H., Fidanoglu, A., Ozcelik, S., & Okumus, A. (2022). User acceptance of autonomous public transport systems: Extended UTAUT2 model. Journal of Public Transportation, 24, 100013. https://doi.org/10.5038/2375-0901.23.1.5 Nordhoff, S., Louw, T., Innamaa, S., Lehtonen, E., Beuster, A., Torrao, G., Bjorvatn, A., Kessel, T., Malin, F., Happee, R., & Merat, N. (2020). Using the UTAUT2 model to explain public acceptance of conditionally automated (L3) cars: A questionnaire study among 9,118 car drivers from eight European countries. Transportation Research Part F: Traffic Psychology and Behaviour, 74, 280–297. https://doi.org/10.1016/j.trf.2020.07.015 OECD. (2017). OECD Guidelines on Measuring Trust. Organisation for Economic Co-operation and Development. https://www.oecd-ilibrary.org/governance/oecd-guidelines-on-measuring-trust_9789264278219-en Pavlou, P. A. (2003). Consumer Acceptance of Electronic Commerce: Integrating Trust and Risk with the Technology Acceptance Model (SSRN Scholarly Paper 2742286). https://papers.ssrn.com/abstract=2742286 Trkman, M., Popovič, A., & Trkman, P. (2023). The roles of privacy concerns and trust in voluntary use of governmental proximity tracing applications. Government Information Quarterly, 40(1), 101787. https://doi.org/10.1016/j.giq.2022.101787

ABSTRACT. The citation based indicators, reflecting the impact of patent on subsequent inventions, are extensively used to measure the quality of patent. At the same time, recent advancement in text mining techniques allows us to look into the novelty of invention by finding new contents which has not found before. A novel or radical invention can be a substantially impactful ones in the future, but it is not always the case. There is a fundamental trade-off between uncertainty in novelty and prominent impact in a process of emerging technology’s development (Rotolo et al, 2015). Understanding such tradeoff is important for STI policy makers in charge of public R&D support for risky projects with great uncertainty but also potential impactful in the future.

This paper address this issues by using our novelty indicator on contextual word embeddings that captures the acquisition or loss of semantics over time, across domains, or cultures by comparing the distribution of the same word across two corpora on a normalized hypersphere (Zhao and Motohashi, 2024). Together with conventional forward citation based impact indicator, we tracks the changes in economic value of a patent throughout the information disclosure process of the patent. Specifically, we investigate the relationship of novelty/impact indicators and stock market valuation of patent information disclosure (application publication and registration) based on the methodology of Kogan et al. (2017), by using 15,418 USPTO patents of 460 US listed pharma firms.

References

Kogan, L., Papanikolaou, D., Seru, A. and Stoffman, N., 2017, Technological Innovation, Resource Allocation, and Growth, The Quarterly Journal of Economics, 132(2), 665-712 Rotolo, D., Hicks, D. and Martin, B.R., 2015. What is an emerging technology?, Research Policy, 44(10), pp.1827-1843 Zhao, Q. and Motohashi, K. (2024), Tracking Semantic Changes with Contextual Embeddings: A Simple, Stable, and Interpretable Method, a paper submitted to AAAI conference, February 2025

ABSTRACT. This paper contributes to the literature on the assessment of transformation policies by introducing a new methodology and indicator framework to better measure scientific novelty. I explore the potential of harnessing advancements in Artificial Intelligence (AI) to develop a new concept – the novelty vector – by delving into unexplored mathematical properties of AI. The novelty vector proved to be an excellent student of novelty scores, learning from expert assessments and predicting high novelty scores with over 90% accuracy.

Methodologically, the approach consists of learning from expert assessment, instead of relying on simplistic assumptions or crude proxies. In an econometric model, the explanatory power of the “novelty vector" retains statistical significance after accounting for the quality of the proposals and applicants’ CVs. When applied to a different dataset of published papers, the “novelty vector" presents equally high levels of predictive power and is statistically significant when identifying Nobel prize-winning papers after accounting for scientific quality. This finding opens a new avenue for further understanding novelty in science and also enables the development of new tools to support policy evaluation in novelty assessment and detection.

Lack of risk-taking in scientific research is weighing on our capacity to generate breakthrough discoveries. We are in desperate need of highly novel and transformative discoveries that can help us tackle grand societal challenges like climate change, ageing, and global pandemics and boost productivity. Despite this need, existing tools to effectively inform, monitor and evaluate the capacity of policies to foster transformative research are notably limited. The lack of robust indicator frameworks for transformative research gives traditional bibliometric indicators an excessive prominence which makes matters worse. Traditional bibliometric indicators based on impact-factor metrics have been shown to be biased against highly novel, transformative research [see e.g. Wang et. Al (2017), OECD (2021) and Machado (2021)].

The current lack of accurate indicators to measure the novelty dimensions of science weighs on our capacity to promote transformational research projects and to monitor and evaluate science funding policies in their capacity to promote disruptive and high-risk/high-gain research. Instead of relying on artificial proxies of novelty defined by bibliometric and scientometrics experts with little knowledge of other scientific domains, this paper proposes a novel methodology that learns from expert reviews in different fields assessing the levels of novelty of scientific ideas in their domains of expertise.

This paper contributes to the scholarly literature dedicated to refining our understanding of the intricate nature of scientific novelty. Franzoni and Stephan (2023) acknowledge how challenging this research area is, encouraging others to take up this important topic. In this paper, I do so by investigating new avenues capable of measuring and predicting how novel new scientific ideas are. This endeavour is not just academic; it holds potential practical significance for science policy. Tools that help identify highly novel ideas have substantial potential to enhance our understanding of the factors that drive novelty and transformational research. Furthermore, this research paves the way for developing innovative tools designed to support peer review processes, streamline the detection of potentially highly novel ideas and help science funders monitor the novelty levels of their portfolio of investments. Such advancement can benefit the broader scientific landscape at several stages, such as project selection, publication processes, and science policy monitoring and evaluation.

The major innovation of this paper is the development of a “novelty vector”. Consequently, the central hypothesis for testing is whether the vector manages to learn from expert reviews and predict novelty using a unique dataset of successful and unsuccessful scientific proposals and their expert review scores. This vector stems from a new type of data exploration using textual embeddings. Embeddings are a form of textual representation originating from machine learning and natural language processing disciplines, which captures semantic meaning based on the context in which textual items appear in large corpora. Embeddings translate words, phrases or documents into vectors of real numbers. These vectors fit a multi-dimensional space such that documents with similar meanings or contexts point at similar directions, while those with distinct meanings are further apart.

Wang, J., Veugelers, R., & Stephan, P. (2017). Bias against novelty in science: A cautionary tale for users of bibliometric indicators. Research Policy, 46(8), 1416–1436. https://doi.org/10.1016/J.RESPOL.2017.06.006 OECD (2021), "Effective policies to foster high-risk/high-reward research", OECD Science, Technology and Industry Policy Papers, No. 112, OECD Publishing, Paris, https://doi.org/10.1787/06913b3b-en. Machado, D. (2021), "Quantitative indicators for high-risk/high-reward research", OECD Science, Technology and Industry Working Papers, No. 2021/07, OECD Publishing, Paris, https://doi.org/10.1787/675cbef6-en. Franzoni, C., & Stephan, P. (2023). Uncertainty and risk-taking in science: Meaning, measurement and management in peer review of research proposals. Research Policy, 52(3), 104706. https://doi.org/10.1016/J.RESPOL.2022.104706

ABSTRACT. Background and rationale

In cases of schemes with multiple aims and multiple assessment criteria, process evaluations face the challenge of having to assess to what extent application selection processes consider and reward these various aspects of proposed research projects. Moreover, recent years have seen a proliferation of modifications to the long-established standard assessment process for research grants (typically external peer review followed by expert panel review using standard criteria around research quality and feasibility) (Sivertsen & Langfeldt, 2025). These include short pre-proposals, inclusion of non-academic reviewers or panellists, in-person presentations. The Austrian FWF’s Emerging Fields (EF) programme aims to fund collaborative research teams to conduct pioneering basic research that departs from established approaches. It aims to give researchers the opportunity to pursue particularly innovative, original, or high-risk ideas. One of the problems with peer review is that proposals based on established knowledge and methods tend to do better than those that include pathbreaking new methods often associated with risk (but potentially also higher rewards) (Langfeldt, 2006). There is evidence that reviewers tend to favour proposals in their fields that align with their ways of thinking, and the more interdisciplinary a grant proposal is, the slimmer its chance of success (Bromham, 2016). Peer review can be biased towards well-recognised names (Jones, 2022) and institutions; there is also some evidence of gender bias (Mutz et al, 2015). The study had to identify whether the EF programme review processes suffer from these biases.

The EF scheme has two separate written application assessment stages: a short outline-proposal stage, followed by a full application review stage. Background research on the scheme’s design led us to a hypothesis that reviews in the first of these two stages should emphasise and reward innovative potential and novelty of the proposed project ideas, while reviews in the second stage should place a greater emphasis on scientific quality of the research plans. Prominent international scientists reviewed the stage-1 synopses and stage-2 full proposals for the EF programme and produced evaluation documents detailing their judgements. In total, we had access to 140 records: 87 peer-reviews of synopses and 53 reviews of full proposals. Given the number, heterogeneity and complexity of the review documents, generative AI was particularly useful in facilitating a systematic assessment (Kolarz et al, 2024).

Data and methods

We used generative AI to analyse peer reviewers’ reports on applications submitted to the FWF’s Emerging Fields (EF) programme. We used a rich and comprehensive generative AI model that managed to navigate these technical details to find relevant individual insights about each review document and stylised facts about the selection process in its two stages. The model was OPENAI's latest GPT4 large language model accessed programmatically via a dedicated API. We performed topic detection, sentiment and priority detection analyses, exploring the text of the review documents from both the first and second assessment stages with generative AI. The topic detection consisted of distinguishing the parts of the reviews focusing on novelty, risk, scientific quality and team suitability dimensions. The sentiment analysis assessed whether the reviewers were positive, negative or neutral on their considerations about each topic. Finally, the priority detection measured the text length or number of words dedicated to each topic as a proxy for how much effort/time the reviewers devoted to each dimension.

Findings

Our results provide evidence in favour of our main hypotheses, most notably with novelty having a significantly higher priority at stage-1 and scientific quality having a significantly higher priority at stage-2. The analysis of priorities confirms that at the synopses review stage, reviewers focus more on the novelty dimension and less on considerations about scientific quality. Our ranking indicator ranges from 0 to 3, with novelty scoring, on average, 2.26 for stage-1 synopses and only 1.25 for stage-2 full proposals. The ‘risk’ and ‘team’ dimensions also receive more attention at the synopses stage than proposals, but with less pronounced differences. The differences in terms of dedication to reviewing scientific quality are more noticeable. This dimension has the lowest rank at stage-1, scoring on average 0.61. In contrast, scientific quality considerations rank the highest at stage-2, scoring 2.19. The findings align with other methods used in the study. Jury members we interviewed frequently commented on the high level of novelty of the synopses they reviewed (and subsequently on the presentations at the Stage 3 Jury hearings). Of our Stage 2 peer reviewers surveyed, around 60% judge the EF application they reviewed to have a higher level of novelty than what they would consider the norm, splitting roughly equally between the ‘slightly’ and ‘significantly’ answer options.

Significance

As documented (OECD, 2021), one central challenge science policymakers face is applicants’ fear that their ideas are too ambitious and risky. Scientists’ fear of following risky ideas is the initial challenge in promoting breakthrough science because, regardless of other potential sources of bias against novelty, breakthrough discoveries are unlikely to emerge without a pool of novel scientific ideas to choose from. Our results suggest that the EF programme was able to convince the scientific community to think outside of the box and submit highly novel ideas for funding.

The methodological significance of the work lays in the use of generative AI to add to the mix of evidence and triangulate across the sources to provide more insightful, comprehensive and robust findings.

References

Bromham L., Dinnage R., Hua X. (2016). Interdisciplinary research has consistently lower funding success. Nature 534 (7609), 684–687. G. Sivertsen, L. Langfeldt (eds.) (2025). Challenges in Research Policy, SpringerBriefs in Political Science, https://doi.org/10.1007/978-3-031-69580-3_5 Jones N. (2022) Authors' names have 'astonishing' influence on peer reviewers. doi: 10.1038/d41586-022-03256-9. Epub ahead of print. PMID: 36220906.

Kolarz, P., Vingre, A., Machado, D., Sutinen, L., Dudenbostel, T., & Arnold, E. (2024). Accompanying process evaluation of FWF's Emerging Fields. Zenodo. https://doi.org/10.5281/zenodo.13911479

Langfeldt, L. (2006). The policy challenges of peer review: managing bias, conflict of interests and interdisciplinary assessments. Research evaluation, 15(1), 31-41.

OECD (2021), "Effective policies to foster high-risk/high-reward research", OECD Science, Technology and Industry Policy Papers, No. 112, OECD Publishing, Paris, https://doi.org/10.1787/06913b3b-en.

ABSTRACT. 1. Introduction Public research funding plays a crucial role in building human capital and scientific infrastructure for national development. In 2022, around USD 2.47 trillion was invested globally in research and development (R&D), with Europe and the US accounting for a significant share. In contrast, Latin America’s R&D spending is approximately 0.7% of GDP, with Colombia investing only about 0.2%. Despite extensive literature on research funding, which has explored funding mechanisms, decision-making processes, and the impact on scientific output, most of this research is focused on and produced by authors from the global north. This focus often overlooks the specific challenges and successes in public research funding in Global South countries like Colombia. This study seeks to enhance the understanding of public research funding in Colombia, especially focusing on the role of matching funds from higher education institutions (HigherEd). We also integrate the use of official open access datasets alongside canonical databases such as Scopus and Web of Science. Our goal is to identify trends in public funding and matching funds in both public and private HigherEd institutions in Colombia between 2013 and 2021 and examine their relationship with institutional research capacities and outputs. 2. Materials We used a dataset from the Ministry of Science, Technology, and Innovation of Colombia (MinCiencias), covering projects and institutions funded by the government from 2007 to 2021. This dataset was cleaned and merged to form a complete table containing information on funded projects, grantee institutions, government transfers, and matching funds provided by the institutions. We also analyzed research groups and knowledge products, focusing on top research groups (ranked A1), which have the capacity to generate high-impact research, engage in international collaboration, and foster innovation. High-quality knowledge products, such as research articles published in top-ranked journals, were also examined as indicators of research capacity. 3. Results We analyzed data from 98 HigherEd institutions—33 public and 65 private—that received funding for at least one project between 2013 and 2020, encompassing 1,423 funded projects. During this period, the government allocated approximately USD 115 million (in 2010-based constant prices) to these projects, with 63% granted to public universities. The average government transfer per project was about USD 98,000, while the total average funding, including matching funds, was USD 205,000 per project. Public institutions housed an average of 271 A1 research groups annually, compared to 159 in private institutions. Additionally, public institutions produced approximately 3,700 high-quality knowledge products annually, while private institutions produced around 2,400. Colombia’s national output of research articles in Scopus-indexed journals grew at an annual rate between 8% and 17%, with an average of 8,500 articles published annually. There were significant fluctuations in total funding during this period, with peaks in 2014 and 2018 and a notable decline between 2015 and 2017, reaching a low in 2019 before a small recovery in 2020. On average, MinCiencias provided around USD 7 million annually in direct transfers, while matching funds from the institutions amounted to approximately USD 7.9 million. This means that matching funds contributed by public and private institutions ranged between 105% and 113% of the direct government transfers, often surpassing the public funding. Despite fluctuations in funding, the number of A1 research groups showed a consistent upward trend. Both public and private institutions experienced growth in the number of top research groups and knowledge products, even during periods of reduced public funding. Public institutions saw a 12% annual growth rate in research outputs, while private institutions exhibited a 16% growth rate. This suggests that institutions may have diversified their financial sources to sustain research activities, such as increasing enrollments or securing public scholarships. Other factors, such as international collaboration, could also influence national research output, though MinCiencias remains the primary national funder. Colombia’s international collaboration in research has seen improvements following a decade of poor performance. 4. Conclusions This analysis aimed to identify trends in public funding and matching funds from both public and private HigherEd institutions in Colombia and their relationship with institutional research capacity and output. Despite fluctuations in funding, there was a consistent upward trend in the number of A1 research groups and top knowledge products, highlighting the importance of matching funds. Institutions in both the public and private sectors have demonstrated the ability to sustain research activities through a combination of public funding and self-generated resources. Future research should incorporate data on private and international funding sources to provide a more comprehensive view of the funding landscape. Additionally, examining the qualitative impacts of funding on broader knowledge production would provide valuable insights into the outcomes of research investments. Comparative studies with other middle- and low-income countries could offer lessons for improving research capacity in similar contexts, enriching the understanding of science policy in developing regions. In conclusion, this study underscores the critical role of matching funds in Colombia’s research ecosystem and suggests that a diversified funding strategy is key to sustaining research capacity, even in the face of fluctuating public funding. By exploring different funding mechanisms and sources, Colombia can continue to enhance its scientific output and contribute to the global knowledge economy.

ABSTRACT. Background While funding is a critical driver of scientific progress, it remains unknown whether and to what extent certain scientific topics are unfunded. There is increasing interest in making the historically hidden scientific funding process more transparent [1], but research is limited because data about unfunded grant proposals is difficult to get access to. The few studies that analyze unfunded proposals focus on the U.S., e.g. [2], and tend to analyze a single cross-section in time, rather than comparing how unfunded topics at an agency change over time. Russia in particular is an interesting case to study how politics and national strategies may influence scientific funding given the identifiable sociopolitical changes over the past 30 years, offering unique insights relevant for understanding the political nature of scientific funding worldwide.

Data & Methods Here, we introduce a new dataset of 326,661 supported and rejected grant applications submitted to the Russian Foundation for Basic Research (RFBR) from 1994-2016, spanning eight disciplines (Biology & Medical Sciences, Chemistry & Material Sciences, Earth Sciences, Engineering, Humanities & Social Sciences, Information Technology, Math & Computer Science, and Physics & Astronomy). The RFBR represents roughly 50% of government funding for science in Russia. Our dataset is one of the largest and most comprehensive of its kind, allowing us to advance the study of the scientific funding process in two main ways: (i) by including rejected applications, and (ii) by analyzing funding dynamics over time. Combined, these contributions expand our understanding about what knowledge might be suppressed by the funding process and characterize the tension between scientific autonomy and governmental agenda-setting.

We constructed the dataset using web-scraping and Russian language assistance from an expert. We scraped each submitted project to the RFBR including the year, title, field, competition, and status (accepted or rejected). We removed duplicates, projects with missing data, and competitions that did not directly fund research (e.g., a competition for conference participation).

Inequities in funding outcomes likely vary by gender and academic experience. In our dataset, we estimated the gender of each applicant using algorithmic name-based gender associations. We estimated the academic experience of each applicant from information about the funding competitions by age (max) and number of publications (min). For example, one early-career competition required applicants to be under age 35 and have more than two recent publications. Our dataset has suitable statistical power to assess differences between subgroups (e.g., women in a given field and year who were funded versus men).

Preliminary Results Scientific understanding of how funding agencies influence what science gets done has been limited by difficulties in accessing funding data, particularly unfunded proposals. This project helps overcome this limitation and greatly advances our understanding of which scientific topics are funded by (i) characterizing the disciplines of funded and unfunded grant proposals, (ii) comparing how over- and under-funded disciplines vary across researchers with different characteristics (gender and academic experience), and at several levels of analysis: overall, cross-field, and within-field, and (iii) identifying how political changes over time impact the disciplines that are funded or not.

Our dataset is uniquely posed to answer these questions because it includes rejected proposals over a long period of time. We analyze time at two separate levels—annually and by sociopolitical period. We constructed three distinct time periods to assess how funding has changed with the cyclical patterns of governmental control in Russia: (i) 1994-2004: Relatively weak state control and low funding, (ii) 2004-2014: Relatively weak state control and explosive growth of funding, and (iii) 2014-2016: Increased state control and continued high funding.

We find substantial gender differences in funded versus unfunded proposals, which vary by discipline. In fact, in almost every field, women’s grants are significantly less likely to be accepted than men’s, with the largest gender gap in information technology proposals. Further, the gender gap in most fields has been growing over time. This could be in part due to shifting state funding priorities over time, with physics and astronomy funding decreasing and engineering and information technology funding increasing starting in 2005, concurrent with the sudden growth of funding in Russia. Finally, we found that women in all fields were more likely to both apply to and receive early-career low-experience awards, however, in Russia these grants are smaller and less prestigious than other grants.

These descriptive statistics alone are powerful given the dearth of data on rejected proposals. But are there differences in the specific research topics that are rejected over time within each field? To find out, we use state of the art topic modeling approaches to characterize the topics among supported versus rejected proposals. We keep the text in Russian and use BERTopic Russian fine-tuned models. Our preliminary results show that the topics that are supported versus rejected differ in the language structure they use across fields.

Significance This project develops an underexplored area in science policy and provides a foundation for policymakers and funding agencies globally to interrogate and improve their practices. For instance, recognition of consistently underfunded topics would impact priority setting at funding agencies. Understanding what science is being proposed but isn’t being done due to the structure of funding will also have implications for the public and researchers. Identifying topics of societal importance that are systematically unfunded could increase socially relevant scientific discoveries. Researchers would benefit from increased transparency in the funding process and greater attention to diversity and fairness. Finally, releasing this new dataset will be extremely valuable for scientists and policymakers interested in understanding and improving scientific funding processes. These results will provide new frameworks to advance research in science and innovation policy, sociology of science, and political science. This abstract demonstrates the analyses that are possible with our data, opening up new directions for future research and new hypotheses.

References 1. Horbach, Tijdink & Bouter. (2022). Research funders should be more transparent: A plea for open applications. Royal Society Open Science, 9(10), 220750. 2. Taffe & Gilpin. (2021). Racial inequity in grant funding from the US National Institutes of Health. Elife, 10, e65697.

ABSTRACT. Research Background While the influence of New Public Management (NPM) has waned in many areas of public administration, its principles continue to dominate research funding allocation across many countries. This persistent influence is evident in several trends: the growing emphasis on competitive project-based funding over institutional block grants, the increasing use of quantitative performance metrics in funding decisions, and the widespread adoption of market-like mechanisms in research resource allocation. The NPM doctrine has fundamentally shaped how research is funded, as particularly visible in the proliferation of performance-based funding systems and the growing administrative burden of grant applications and reporting requirements. This continued dominance of NPM principles in research funding has become especially problematic in the era of platform-based and networked science. Modern scientific research increasingly relies on large-scale research infrastructure, real-time data sharing, and complex collaborative networks that transcend institutional and national boundaries. These characteristics fundamentally conflict with NPM's emphasis on competition, clearly defined deliverables, and short-term performance metrics. This misalignment discourages risk-taking in emerging fields, hampers the development of sustainable research infrastructure, and creates barriers to fluid, cross-institutional collaboration necessary for advancing contemporary science. Research Questions This study addresses five interconnected research questions. First, how do platform-based and networked characteristics of modern science challenge the NPM-influenced balance among different research funding mechanisms? Second, what institutional factors explain the persistence of NPM principles in research funding despite their apparent limitations? Third, how do varying levels of governance quality affect the optimization of funding mechanisms across different countries? Fourth, what innovative approaches are emerging to reconcile accountability requirements with the needs of networked research? Finally, how can funding mechanisms be recalibrated to better support platform-based science while maintaining necessary oversight? Research Methods This study develops an analytical framework integrating institutional theory, transaction cost economics, and innovation systems theory to examine how funding mechanisms align with demands of platform-based and networked science. The framework analyzes three key dimensions: funding mechanism balance (distribution among block grants, individual funding, and project funding), governance quality (transparency, efficiency, and accountability), and platform/network characteristics (infrastructure requirements, collaboration patterns, and data sharing needs). The research employs a comparative case analysis of five major research economies (the United States, Germany, Japan, the United Kingdom, and China), selected for their varying balances of funding mechanisms and different approaches to research governance. Our analytical strategy combines four methods. First, historical analysis traces the evolution of funding mechanisms (2000-2024) through systematic documentary analysis of national research funding policies, funding agency annual reports, program documentation, and reform proposals. Second, quantitative analysis examines funding allocation patterns, including the distribution across mechanisms, trends in allocation, and relationships with research outputs. Third, institutional analysis investigates governance structures, funding program designs, and evaluation systems through detailed examination of institutional documents and program guidelines. Fourth, cross-national comparison systematically examines variations in funding approaches and their outcomes across different national contexts. Data collection focuses on three main sources. The primary dataset comprises documentary evidence including national research strategies, funding agency policies, program evaluations, parliamentary/congressional reports, and institutional guidelines. This is supplemented by quantitative data on funding allocations, research outputs (publications, patents, collaborative projects), and infrastructure investments. Additionally, selective expert interviews will help validate and interpret findings from documentary and quantitative analysis. The data will be analyzed using a comparative matrix examining relationships between funding mechanisms, governance structures, and research outcomes across different national contexts. Expected Findings Drawing from our integrated analytical framework, the comparative analysis is expected to reveal several key findings. First, institutional theory would help explain why countries with strong governance structures but different institutional traditions (like the US and Germany) have evolved distinct hybrid funding models that depart from pure NPM principles. Second, transaction cost analysis is likely to demonstrate how different types of research platforms and networks require distinct funding configurations - with research involving shared infrastructure and extensive collaboration showing higher transaction costs under traditional competitive funding. Third, innovation systems analysis would reveal how the effectiveness of funding mechanisms depends on the alignment between governance arrangements and research infrastructure requirements, with better-aligned systems demonstrating higher innovation outputs and stronger collaborative networks. These findings collectively suggest that the optimal balance of funding mechanisms varies with institutional contexts and research characteristics. Theoretical and Practical Contributions This research advances both theoretical understanding and practical knowledge of research funding mechanisms. Theoretically, it extends institutional and transaction cost theories by explaining how platform-based and networked science creates new organizational demands that challenge traditional NPM principles. The integrated analytical framework developed in this study provides a novel approach for analyzing the alignment between funding mechanisms and research characteristics in different institutional contexts. Practically, this research offers evidence-based guidance for policymakers on optimizing the balance between funding mechanisms based on institutional conditions, governance capacity, and research infrastructure requirements. These insights can help funding bodies design more effective funding portfolios that support contemporary scientific collaboration while maintaining appropriate accountability measures.

ABSTRACT. Background and research questions Germany and Europe face significant transformations. The automotive industry struggles to compete in the electric vehicle sector, while the chemical and steel industries demand clarity on energy prices and innovation funding. Additionally, German businesses must enhance their competitiveness in artificial intelligence applications. Despite the ecological transformation often being viewed as a burden, green technologies are essential for long-term competitiveness and economic prosperity. Relying on traditional technologies, such as combustion engines, risks an "innovation dilemma," securing short-term profits but losing out in the long run. A proactive industrial policy is increasingly demanded to ensure prosperity, competitive-ness, and decarbonization amid technological and geopolitical shifts. This policy must balance horizontal measures, like energy pricing, with vertical support for specific sec-tors. Absolute technological neutrality is neither feasible nor desirable. Early-stage open-ness to diverse technologies should eventually lead to clear prioritization to align infra-structure and investment. Germany must also respond to global shifts, such as the U.S. Inflation Reduction Act, by defining strategic priorities and fostering key sectors. Facing these challenges of industrial policy making in Germany and Europe, recent dis-cussions on dynamic capabilities of the state towards mission-orientation highlight state capacities as necessary condition. But what kind of capacities governments and admin-istrations need to learn and use for industrial policy making? While the necessity of dy-namic capabilities, strategic intelligence and anticipatory governance ae acknowledged by an increasing number of states, international organizations and public administrations, questions emerge in regard to their concrete institutionalization. Current research on institutionalizing strategic foresight shows how to analyze and support the adaption of forward-looking reflexive policy practices(Priebe et al. 2024). Despite clear links to the challenges in industrial policymaking, there is a surprising lack of research on the appli-cation of foresight in this policy domain, notably for the translation of industrial strategies to an (adaptive) industrial policy mix. Thus, we are rising the research question on what kind of state capacities governments and administrations need to conduct industrial poli-cies and how strategic foresight contributes to that.

Methodology By reviewing the literature on industrial policy making for strategies and mixes of horizon-tal and vertical instruments, we deduced the main principles to conduct a rational indus-trial policy. Moreover, based on the foresight literature on technological and strategic foresight, we created a foresight process for the analysis of scenarios, the deduction of industrial policy objectives and the choice and design of industrial policy instruments. We are currently starting several workshops with policy makers at the federal or federal state level in Germany to test this process and to learn how from a practitioners’ perspective industrial policy making need to be discussed and conceptualized.

Literature Review: Principles for effective industrial policy Industrial policy thrives on clear goals and societal integration, balancing broad participa-tion for practical solutions with the risk of excessive lobbying. For example, transforming heating systems requires combining societal input with expert guidance. While industries like automotive benefit from top-down directives (e.g., prioritizing electromobility), sectors like heating need participatory approaches to identify viable technologies for varied needs. Clear direction is vital. The state must provide coherent signals about future markets and phase-outs, such as transitioning to renewables or vehicle electrification. Credibility relies on aligning ambitious targets (e.g., renewable energy) with necessary infrastructure, while addressing urban and rural differences like charging stations or district heating. A balanced mix of financial, regulatory, and informational instruments is crucial. Financial incentives foster innovation by reducing risks, while regulations establish clear market rules. Informational measures, like campaigns or data collection, address knowledge gaps, as seen in the heating transition. Local and regional dynamics require collaboration across governance levels, from municipal planning to European energy integration. Timing and flexibility are key. Industrial policy must anticipate technological and geopolit-ical shifts and allow adjustments. For example, delayed photovoltaic subsidy reforms caused high costs, highlighting the need for adaptable, scenario-based planning. Policies must ensure readiness, enabling households and industries to adopt sustainable tech-nologies without excessive strain. Finally, accountability ensures legitimacy. Clear, measurable goals and transparent pro-gress reviews foster adaptability, support public acceptance, and enhance policy evolu-tion.

Preliminary results: Design of a strategic foresight process for industrial policy Industrial policy must identify the key core sectors of Germany’s current economy, de-fined by their significant contribution to GDP, employment, or regional economic struc-ture. Future sectors extend these core sectors over a 20-year horizon, potentially evolv-ing from existing ones or emerging as new fields, such as the hydrogen economy. Addi-tionally, strategic sectors, like medical protective equipment, defense, semiconductors, or other critical intermediate goods, are vital for geopolitical reasons, even if they contribute less to GDP. Ensuring sufficient production of these sectors in Germany or Europe is a strategic necessity. Identifying relevant sectors involves significant uncertainty. The evolution of technolo-gies, socio-economic conditions, and geopolitical factors determines which sectors are considered future or strategic. Strategic foresight analyzes these possibilities through internal workshops or broad participatory processes. Society must adapt to various sce-narios and understand industrial policy adjustments. Scenarios outline potential futures based on key influencing factors, such as geopolitical conflicts, EU cohesion, technological developments, or shifting societal priorities. Once developed collaboratively, these scenarios guide discussions on which sectors are rele-vant in different contexts. They also inform industrial policy goals. If a future sector ap-pears in all scenarios, robust strategies should focus on its promotion as a "no regret" option. Conversely, sectors relevant in specific scenarios should be supported with flexible policies accommodating changing priorities. From these largely narrative scenarios, indicators can be derived to analyze which scenario aligns with current realities. Regular monitoring helps implement scenario-specific strategies, prioritize instruments, and adjust goals as needed. If trends indicate a shift toward another scenario, policymakers can preemptively evaluate affected sectors, abandon outdated goals, and prioritize new measures. By integrating foresight, monitoring, and flexibility, industrial policy can navigate uncer-tainty, foster innovation, and maintain societal and economic resilience, aligning strategies with evolving geopolitical, technological, and ecological landscapes.

ABSTRACT. Innovation is the foundation of economic development and societal advancement. It enhances productivity and competitiveness, while simultaneously addressing critical global challenges such as resource scarcity, health crises, and climate change. However, the process of fostering innovation is intricate, involving policies and context. While existing research has found that informal competition impacts product innovation by formal firms in developing countries, there is a lack of understanding of how this impact may change or may be moderated depending on the institutional environment quality. This study aims to examine the role of institutional quality on the relationship between informal competition and formal firm innovation. We contribute to the literature in three ways. First, motivated by the “legalist” theory of informality, our study highlights the role of institutional quality in the relationship between informal competition and formal firm innovation. Second, complementing recent findings on informal competition and firm innovation nexus, we examine this relationship considering innovation performance relative to both input (resources involved) and output (results obtained). We pay special attention to the multi-dimensionality of innovation instead of treating it as a homogenous construct. Third, another contribution of this work lies in the consideration of the non-linearity effect of informal competition on formal firm innovation. This empirical contribution allows us to test the “Schumpeterian effect” vs. the “Arrow effect” or the Aghion hypothesis. The data source used in this study is the World Bank Enterprise Survey (WBES). These data cover the period 2011-2022 and include 6,979 formal manufacturing firms from 19 sub-Saharan Africa countries. Our empirical strategy is based on multivariate probit model, that taking into account the dependence between the dependent variables: R&D, product innovation and process innovation. Using the average values of competition and institutional quality in each country allows us to overcome the problem of endogeneity of the explanatory variables of interest. Our results show that the marginal effect of informal competition on product and process innovation by formal firms increases and then decreases with the intensity of competitive pressure. This result reconciles the divergent perspectives of Schumpeter and Arrow, establishing an inverted-U relationship between competitive pressure from informal firms and propensity to introduce new product and process in the market. In other words, informal competition promotes output innovation in formal firms as long as it remains moderate. Beyond that, it becomes harmful to innovation. However, input innovation is linear and decreases with the intensity of informal competition. Moreover, the results show that the quality of institutions moderates the effect of informal competition only in regions where that competition is strong. The most important institutional factors are in this order: the tax rate and tax administration, the court system, corruption and political instability. Our findings have important practical implications for policymakers across SSA countries to enhance innovation in manufacturing sector. Policy makers should regulate the extent of informal sector activity through policies that neither eliminate informal firms entirely nor allow their dominance. Government should provide formal firms with incentives for innovation, such as tax breaks or subsidies, to mitigate the negative effects of high competition. Given that input innovation declines with the intensity of informal competition, targeted policy such as subsidizing R&D in sectors heavily impacted by informal competition is needed. Further, institutions play a critical role in moderating the effects of informal competition, highlighting the importance of investments in governance, law, and tax systems. Supporting innovation should involve fostering a collaborative ecosystem where formal and informal firms coexist productively. These implications provide a framework for concrete strategies to address the dual challenges of informality and institutional quality while fostering firm innovation in sub-Saharan Africa manufacturing sector.

ABSTRACT. Introduction and Background Innovation and incubation centers are pillars of Taiwan's innovation landscape. According to Hsien-Hsiang Lee, they serve as essential drivers of the country's innovation ecosystem. These centers are designed to support startups, small and medium-sized enterprises (SMEs), and foster technology transfer by providing a range of resources such as co-working space, administrative support, technology commercialization services, and financial and legal consulting. Taiwan’s innovation ecosystem is characterized by a strong collaboration between academia, industry, and government, which has been instrumental in driving its global competitiveness and advancing technological innovation, particularly in information and communications technology (ICT), agricultural technology (AgriTech), and medical technology (MedTech).

This study investigates the function of innovation and incubation centers and their multifaceted contributions to the dynamic innovation ecosystem. By examining their structure, functions, strategic focuses, and recourses of these centers, as well as understanding stakeholder opinions and perceptions of the innovation environment, this research seeks to provide insights into how these centers shape Taiwan’s innovation landscape.

Methodology The two primary methods employed in this research were a literature review and qualitative, semi-structured interviews to explore the role of innovation and incubation centers within Taiwan’s innovation ecosystem. The study began with a scoping review of existing literature on Taiwan's innovation policies, government initiatives, and academic research related to innovation and incubation centers. This review provided a foundational understanding of the current landscape of innovation centers in Taiwan.

To gain deeper insights into the operations and challenges faced by these centers, we conducted interviews with 15 experts, including center directors, industry partners, and academics from both university-based (n=9) and research institute-based (n=6) centers. Each participant took part in a single 60-minute interview designed to gather information about their organization, personal experiences, and opinions about the present innovation landscape. The interview protocol focused on key themes in governance structures, resource allocation and availability, strategic partnerships, technology transfer processes, and international collaboration.

Results and Discussion One of the key findings indicate that Taiwan’s innovation and incubation centers emphasize collaboration among academia, industry, and government to strengthen the innovation ecosystem. Drawing from insights provided by the Small and Medium Enterprise and Startup Administration (SMEA) under the Ministry of Economic Affairs, as well as participant perspectives, university-based centers leverage their access to entrepreneurship and innovation education (EIE) programs, research facilities, and faculty expertise to assist startup teams and SMEs in commercializing research outputs, which effectively bridges the gap between theory and practice. Research institute-based centers, such as those in Hsinchu Science Park, demonstrate co-creation models where high-tech companies like Taiwan Semiconductor Manufacturing Company (TSMC) collaborate with local startups and universities to develop cutting-edge technologies. These partnerships and cross-sector collaborations are instrumental in creating integrated industry supply chains and fostering an open, innovative, and collaborative environment.

In addition, these centers aim to support early-stage startups and SMEs. They provide resources, including operational space, access to research facilities, mentorship, support in sourcing funding, and business consulting services such as legal and financial advisory assistance. These efforts are important in Taiwan’s economy, where over 98% of businesses are SMEs, according to Taiwan’s White Paper on Small and Medium Enterprises (2022). For example, university-based centers focus on reducing costs for startups by offering subsidized access to infrastructure and expertise. Similarly, research institute-based centers provide SMEs with market intelligence and technical support to accelerate product development cycles.

Despite these strengths, limited flexibility in fund utilization and insufficient manpower remain barriers to fully supporting entrepreneurial growth and meeting their needs. Many innovation centers mention challenges with defining a clear role within their institutions and the broader innovation landscape. Due to limited budgets and resources, these centers often diversify their services to attract startups and external collaborators. Although this approach increases engagement, it can dilute the center’s core mission, leading to inefficiencies in resource utilization and unclear positioning and limited effective interaction between center members and resident companies.

On the other hand, experts share challenges in securing sustainable long-term funding without over-reliance on government subsidies or large corporations. For instance, while Hsinchu Science Park business model has achieved global recognition for its semiconductor ecosystem through strong supply chains and government-industry-academia partnerships, smaller centers in other fields suggest difficulties in replicating this pattern due to limited financial resources. A shift toward diversified funding models—including private investments and international grants—could help mitigate these challenges.

Last but not least, Taiwan has made strides in internationalizing its innovation ecosystem, there is still room for improvement in fostering deeper cross-border collaborations. Participants observe that Taiwanese startup teams and SMEs tend to lack familiarity with foreign and global markets. Despite the fact that they excel in developing deep-tech solutions, they require additional efforts to effectively showcase their outcomes on the global stage and attract potential investment and collaboration opportunities. Expanding these collaborations could open access to new markets and talent pools, which strengthens Taiwan’s innovation ecosystem and supporting its global competitiveness.

Conclusion The study advances our understanding of how innovation and incubation centers contribute to Taiwan’s innovation ecosystem. The findings suggest that future policy efforts should prioritize clarifying the positioning of innovation centers, systematically organizing available resources to increase funding availability for early-stage startups and SMEs, and fostering international collaborations. Moreover, these findings inform policymakers about providing tailored support and resources to different types of centers. By recognizing their missions and operational needs, policymakers can corroborate that both are adequately resourced to facilitate their contributions to the innovation ecosystem.

ABSTRACT. The prevailing emphasis on technological innovation often overshadows the critical importance of maintenance and repair in improving infrastructure reliability and sustainability. This innovation-centric mindset, as critiqued by maintenance theorists, leads to systemic vulnerabilities in technological systems due to the undervaluation of upkeep activities. In the context of electric vehicle (EV) charging infrastructure, this oversight manifests as declining operational reliability and quality, despite significant investments aimed at expanding the network to combat climate change. This study examines the deteriorating reliability of EV charging stations in the United States amidst substantial federal investments, such as the Biden administration’s $7.5 billion allocation for building a national charging network. While the expansion of charging infrastructure is crucial for promoting EV adoption and reducing carbon emissions, the maintenance of these stations remains inadequately addressed. The reliability issues not only hinder user experience but also pose a significant barrier to EV adoption due to indirect network effects, where the value of the network depends on the reliability of its components. Drawing on the theoretical framework that highlights the overlooked significance of maintenance, this research investigates how nearby service density influences the operational reliability of EV charging stations. Specifically, it explores two competing mechanisms: 1. Competition Incentivizing Maintenance: Greater competition from nearby stations may incentivize hosts -- especially those generating revenue directly from charging services -- to maintain high reliability to attract and retain customers. This aligns with the notion that when maintenance directly impacts revenue, it becomes a prioritized activity. 2. Free-Riding Leading to Maintenance Neglect: Conversely, an increased density of nearby stations may lead some hosts, particularly those for whom charging is a supplementary service, to rely on others to provide reliable service. This results in underinvestment in maintenance due to the diffusion of accountability -- a phenomenon observed in broader maintenance neglect patterns. Existing research on competition and free riding primarily focuses on the service providers within the same industry offering the same core service. There is a lack of studies examining competition in the context of core versus supplementary services, where providers of the same primary business are differentiated from those of different primary businesses. The study formulates two hypotheses to test these mechanisms: H1: An increase in the density of nearby same-industry service providers leads to higher station reliability, especially for core service providers. H2: An increase in the density of nearby different-industry service providers leads to lower station reliability for supplementary service providers. To empirically investigate these hypotheses, a comprehensive panel dataset of U.S. charging stations from 2011 to 2024 was compiled from multiple sources. The nearby service density was measured as the number of stations within a 1-mile radius of each station, with robustness checks conducted using radii from 0.2 to 5 with a step of 0.2. Charging reliability was quantified by analyzing consumer reviews from EV charging locator apps, focusing on the percentage of reviews mentioning reliability issues. Advanced natural language processing techniques, including fine-tuned BERT and GPT models combined with chain-of-thought prompting, were employed to classify relevant user feedback accurately. A fixed-effect lagged panel regression model was used to control for unobserved heterogeneity and to capture the dynamic effects of nearby service density on station reliability over time. The descriptive analysis revealed a concerning trend: the reliability of EV charging stations across the U.S. has been steadily declining each year, falling significantly below the expected 97% uptime mandated by federal policy. The regression results supported the proposed hypotheses. Specifically: H1 confirmed: An increase in the density of nearby same-industry service providers leads to higher station reliability, especially for core service providers. For core stations, a 100% increase in the prior year’s density of nearby stations of the same business type is associated with a 6.93-point increase (10.77% increase) in the reliability score. This suggests that competition incentivizes these hosts to invest in maintenance to attract customers, reinforcing the theoretical perspective that maintenance becomes prioritized when directly tied to revenue generation. H2 confirmed: An increase in the density of nearby different-industry service providers leads to lower station reliability for supplementary service providers. These hosts appeared to engage in free-riding behavior, underinvesting in maintenance with the assumption that nearby stations would compensate for their lack of reliability -- a manifestation of the diffusion of maintenance responsibility. These findings underscore the limitations of relying solely on market forces to ensure the reliability of critical infrastructure. The observed maintenance neglect aligns with the theoretical critique that innovation-centric policies often fail to address the essential work of upkeep, leading to systemic inefficiencies and user dissatisfaction. The study highlights the necessity for policy interventions to address the maintenance gap in EV charging infrastructure. This work demonstrates that the market is under-providing service quality, highlighting the need for policy interventions. Greater funding and incentives should be directed not only towards building new chargers but also towards maintaining and inspecting existing ones. The government must establish clear accountability for reliability standards, such as enforcing a minimum operational threshold (e.g., 97% uptime). In the long run, R&D investment should be encouraged to develop more durable and resilient charging equipment. By integrating the theoretical insights on the undervaluation of maintenance, this study illuminates a critical gap in the current approach to EV infrastructure development. The declining reliability of charging stations, exacerbated by inadequate maintenance practices, poses a significant barrier to EV adoption and undermines efforts to combat climate change. Recognizing maintenance as a fundamental component of technological systems is essential for achieving long-term reliability and fostering user trust in EV charging networks and therefore increasing the adoption of EV and other clean energy products. This research contributes to the broader literature on the importance of maintenance in technological systems and offers practical policy recommendations to enhance the reliability of EV charging infrastructure. Addressing the maintenance gap is imperative not only for the success of EV adoption but also for ensuring the sustainability and resilience of critical technological infrastructures in the face of growing environmental challenges.

ABSTRACT. Public use files (PUF) of federal survey microdata on individuals have been available for decades. However, large increases in computing power and the greater availability of Big Data have dramatically increased the probability of re-identifying anonymized data, potentially violating the pledge of confidentiality given to survey respondents. The same data science tools that increase the risk of disclosure can also be used to produce synthetic data that preserve critical moments of the empirical data but do not contain the records of any existing individual or business respondent. These synthetic data tools open new possibilities for producing microdata that will allow producing informative descriptive or multivariate analyses using PUFs that currently require access to confidential microdata. Developing public use establishment data from surveys presents unique challenges from demographic data, because there is a lack of anonymity and certain industries can be easily identified in a given geographic area. The presentation will briefly describe an algorithm used to construct a synthetic public use file based on the 2019 Annual Business Survey (ABS) and discuss various quality metrics. The ABS is conducted jointly by the Census Bureau and the National Center for Science and Engineering Statistics (NCSES) within the National Science Foundation (NSF), and collects data on R&D expenditures, innovation-related data, globalization, and business owner characteristics from businesses operating in the U.S. Various use cases—either to substitute or supplement accessing confidential data—will be discussed in the context of tiered access.

ABSTRACT. Launched in 2018, STIP Compass is an online interactive data repository that gathers data on national science, technology and innovation (STI) policies. Its overarching goal is to contribute to the monitoring and assessment of STI policies in OECD and European Union (EU) member countries and beyond. It does this through a joint OECD and European Commission (EC) digital infrastructure, which is used to survey countries on their STI policies, link the collected policy data with other information (including statistics and other databases), and provide open data services through a dedicated web portal accessible at http://stip.oecd.org.

A biennial survey on STI policies, branded as the EC-OECD Science, Technology and Innovation Policy (STIP) Survey, constitutes the portal’s original and central dataset. It is unique in its scope, nature and coverage. The survey collects qualitative information on policy initiatives, defined and characterised through several open text-fields (e.g. name, objectives, short description, start/end date, etc.) and through multiple choice fields describing policies’ direct beneficiaries, policy instruments and budget ranges, among other characteristics.

The survey’s topic coverage is structured around six “core” policy areas: (i) Governance, (ii) Public research system, (iii) Innovation in firms and innovative entrepreneurship, (iv) Science-industry knowledge transfer and co-creation, (v) Human resources for research and innovation, and (vi) Research and innovation for society. It also includes survey edition-specific “modules” that cover selected key issues of concern to the EC’s and OECD’s policy agendas and work programmes, such as STI-enabled net zero transitions.

The survey is answered by several hundred government officials, with coordination carried out by national delegates of the OECD’s Committee for Scientific and Technological Policy (CSTP). The 2023 edition of the STIP survey gathered information on more than 7,700 policy initiatives from 59 countries and the EU. The level of reporting and data quality varies by country, with some registering over 200 policies while a few others report less than 50. The average is around 100 policy initiatives per country. The policy initiatives reported vary widely in scope and scale, from ambitious national agendas to smaller R&D grant schemes. The next survey will run in spring 2025.

STIP Compass uses semantic web / linked data technologies to store, link and make accessible the information collected by the survey. These technologies provide much greater flexibility than a traditional relational database and can admit any type of data. They rely on controlled vocabularies to structure data, opening it up for linking to other types of STI data from the EC, OECD and other organisations. These include qualitative data (e.g. metadata on +10,000 EC and OECD publications and +40,000 academic articles), as well as quantitative data accessible through the STI.Scoreboard infrastructure (including +200 unique country statistical indicators, such as expenditures in R&D, doctoral graduation rates, and patent applications). This “linked data” is brought into the same database primarily to contextualise the STI policy data collected in the survey.

While both the OECD and EC make extensive use of the data in their country and thematic analyses, the expected user base is considerably broader and includes national policy makers and analysts (in governments and elsewhere, including academia). This wide user base translates into a broad range of uses of STIP Compass, ranging from consulting facts on national policies to conducting analysis of policy mixes. While accounting for data limitations, the dataset provides a wealth of evidence to map national policy mixes, compare international approaches to STI policy, enable policy learning, and tackle research questions in the field of STI policy. The STIP Compass site attracted around 200,000 unique visits in 2023 and its traffic continues to grow.

To promote data accessibility, STIP Compass has more than 700 interactive dashboards that allow users to browse the database and to zoom in and out of the data. Each dashboard has several interactive panels that show relevant policies and other types of quantitative and qualitative data, such as statistical indicators and publication metadata. As a data interface, they enable users to intuitively and visually navigate the STIP Compass database. Nevertheless, users can be overwhelmed by the wide thematic breadth of these dashboards. To overcome this limitation, STIP Compass has developed thematic portals, which are sections of the site that use tailored interfaces to bring additional layers of guidance and/or additional focus to specific policy issues, societal challenges and business sectors. Topics covered include open science, research security, Indigenous knowledge and communities, and net zero.

The analyst community can also freely download the STIP survey datasets in human- and machine-readable formats for their own analysis. The STIP Compass infrastructure achieves this through an open API, a common way for sharing open data. The download data page has a “query builder” interface that allows any user to query the STIP Compass server and download data from the database. The site also hosts a STIP Data Lab to exploit STIP Compass policy data using new analytical capabilities (e.g. natural language processing) that provide exemplars of the data’s analytical potential. The Lab aims to engage government officials and STI policy experts in co-designing and co-producing analyses, with a view to building a community of STIP Compass “lead users”.

The first part of this presentation will (i) describe and explain how this data infrastructure has been set up and the rationales informing its development; (ii) briefly present the data and features it contains; and (iii) outline how the infrastructure can be used for policy analysis and advice, particularly by the research community.

A second part will focus on recent disruptions that challenge the ways the project has so far been carried out. These include (i) emerging digital technologies, such as LLMs, which affect all phases of the project's data cycle; (ii) shifts in STI policy agendas that increasingly focus on promoting sociotechnical transformations where STI is not expected to play a leading role; and (iii) rising geopolitical tensions that can hinder the maintenance of an international harmonised policy database like STIP Compass.

*Current role: the authors are OECD-based analysts responsible for developing and maintaining STIP Compass and analysing its data.

ABSTRACT. The National Center for Science and Engineering Statistics (NCSES) within the U.S. National Science Foundation (NSF) is the principal source of analytical and statistical reports, data, and related publications that describe and provide insight into the nation’s science and engineering resources. NCSES collects and provides data on the science and engineering workforce; research and development (R&D); competitiveness in science, engineering, technology, and R&D; and the condition and progress of STEM education in the U.S. These data are collected as sample surveys or census.

The State of U.S. Science and Engineering summarizes key indicators that assess the status of the science and engineering (S&E) enterprise within the United States and illustrate the U.S. global position in multiple aspects of the S&E enterprise. This report provides high-level findings from the nine thematic reports that make up Science and Engineering Indicators 2024. Selected data from the nine thematic reports are grouped into three major sections that relate to the S&E enterprise: talent, discovery, and translation. These three components collectively support U.S. global competitiveness in science, technology, and innovation (STI), in that science, technology, engineering, and mathematics (STEM) talent contributes to scientific discovery, which in turn is translated to society and the economy through innovation.

The first section of the report describes the status of the U.S. STEM education system from elementary through the doctoral level and the STEM workforce, including the contributions of international students and workers. It also details the American public’s perceptions about scientists. The second section, on research and development (R&D), discusses the position of the United States among the top R&D-performing countries and analyzes patterns of U.S. R&D funding and performance among economic sectors and by type of R&D. The third and final section focuses on outputs of the S&E enterprise to provide insight into how the United States and other major countries and regions contribute to global knowledge and innovation. Finally, the report contains a sidebar with select indicators of national investments and capabilities in critical and emerging technologies.

Science and Engineering Indicators (Indicators) provides high-quality quantitative information on the U.S. and international science and engineering (S&E) enterprise. Indicators consists of detailed thematic or focus area reports, a state data tool, and a congressionally mandated report delivered biennially to the president and Congress that highlights important trends from across the focus areas. Indicators reports employ a variety of presentation styles—such as narrative text, data tables, and figures—to provide accessible data to consumers with different information needs.

The data described in Indicators are a quantitative summary of the scope, quality, and vitality of the S&E enterprise over time and within a global context. These data are intended to contribute to an understanding of the current environment and to inform the development of future policies.

Indicators is prepared under the guidance of the National Science Board by the National Center for Science and Engineering Statistics (NCSES), a principal federal statistical agency within the National Science Foundation, Social, Behavioral and Economic Sciences Directorate. NCSES develops the content and the dissemination platforms. Indicators reports are subject to extensive review by internal and external subject matter experts, federal agencies, National Science Board members, and NCSES statistical reviewers for accuracy, coverage, and balance.

ABSTRACT. # Introduction Putting science into effective action—whether by informing policy, guiding practice, or planning a new study—requires that individual papers not be considered in isolation, but rather in the context of the wider scientific literature. A key consideration is the criticism a study receives. Healthy critique between academic peers helps to identify errors in the scientific record and forces scholars to reckon with competing evidence and assumptions. Criticism may also signal that a paper’s findings require further supporting evidence before informing policy and practice.

The utility of criticism rests on it actually being read, an assumption that may not actually hold in reality. Retracted papers have been found to accumulate citations and attention long after retraction, likely owing to scholars following heuristic-based citation strategies rather than carefully reading and engaging with the literature. The same may hold true for criticism.

This abstract details results of an ongoing project examining the visibility of criticism in science. I focus on critical letters: a publication type with the purpose of critiquing a recently published paper. Over 3,000 critical letters were sourced from several large, widely-read, and disciplinary-diverse journals in which they are routinely published sourced from the journals Nature, Science, PNAS, and the journals of the American Physical Society (APS). Bibliometric metadata is sourced from the final 2021 snapshot of the Microsoft Academic Graph and includes critical letters published between the years 2000 and 2018.

# Results Visibility is examined along a number of dimensions. The first is based on citations accumulated. Across each journal, papers targeted by a critical letter continue to accumulate citations well after the criticism was issued. While there is a decline in annual citations, this does not appear attributable to the critical letter itself. In itself, this is unsurprising; criticisms are not like retractions, in which citation to the targeted paper which does not acknowledge the retraction is prima facie inappropriate. Other results not shown here suggest that critical letters have no detectable effect on the citations received by a paper. What is surprising is that the critical letters go comparatively uncited. Only a small fraction of citations to the targeted paper are accompanied by a citation to the critical letter—around 10% for elite generalist journals and 15% for APS journals in the initial years, declining over time. I also investigate visibility based on altmetric indicators. The number of tweets and news stories associated with a paper are sourced from the SciSciNet dataset. The results show that critical letters receive considerably fewer tweets and mentions in news stories than those they critique.

The visibility of critical letters is not uniform across all readers. Those authors most familiar with the field of the criticized paper are most likely to encounter the letter and consider it relevant to cite, whereas those readers from more distant fields may see the original paper but not its criticism. To test this, I use SPECTER, a fine-tuned version of the BERT language model, to generate vector representations for each paper targeted by a critical letter as well as for all its citing papers. The cosine distance between these vectors provides an approximate measure of topical similarity. A similar approach has been used previously to evaluate the topical distance between retracted papers and their post-retraction citations. After excluding replies and self-citations, I find that papers co-citing the critical letter tend to be among the most topically similar to the targeted paper. This effect is most pronounced in elite generalist journals such as Nature (μ = 59.3), Science (μ = 58.9) and PNAS (μ = 57.5). For the specialist APS journals, the effect is present but weaker, possibly because these journals primarily attract citations from within their own field, reducing the likelihood of cross-disciplinary citations.

In summary, critical letters appear to have limited visibility. These findings suggest that critical letters are most visible within their immediate disciplinary communities, whereas researchers in more distant fields may be less likely to encounter or cite the critical letter. This pattern implies that the influence of critical letters may be constrained by disciplinary boundaries, with limited impact on the broader academic audience.

# Significance Criticism is widely regarded as a fundamental component of science and is necessary to holistically assess the state of the science on any particular topic. While critique is ubiquitous across all channels of scholarly communication, critical letters represent one of the clearest and most explicit forms of criticism in science. Yet these findings suggest that critical letters are not being widely read nor well incorporated into the academic literature.

Critical letters also provide a useful framework for understanding the implications of post-publication peer review, replications, red-team reports, and other forms of critique and review issued after a paper is published. Namely, such practices may run into similar issues of visibility.

These findings pose a challenge for science. Journals may consider interventions to ensure criticism is not overlooked, ensuring that citation practices reflect a comprehensive understanding of each work’s context. Particular care is needed where science guides policy or practice, such as in systematic review and evidence synthesis. Here, papers should be carefully considered in the context of their critiques.

ABSTRACT. Abstract: This paper investigates the impact of specialization on the career development of PhD students in the life sciences, with a focus on both short-term and long-term effects. It examines how specialization during the PhD period influences staying in academia immediately after graduation and the likelihood of obtaining tenure promotion later in a career. Using data from 685 PhD students in Japan, the paper finds that specialization during the PhD positively influences staying in academia but has an inverted U-shaped effect on tenure promotion. Specialization in later years, however, negatively affects tenure promotion. These findings suggest that while specialization helps PhD students establish expertise early in their careers, over-specialization may hinder long-term academic growth.

Introduction: The paper addresses a gap in the literature by exploring how specialization influences academic careers, particularly for PhD students. Specialization is seen as a double-edged sword—it may enhance research performance by deepening knowledge but also potentially stifle creativity, which is crucial for long-term academic success. Specialization can benefit young researchers by helping them focus and gain expertise quickly, but it may later limit their capacity for creative breakthroughs as they progress in their careers. The study aims to understand how specialization affects academic careers both immediately after PhD completion (short-term) and in the longer run (long-term, including tenure promotion).

Data and Methodology: The study uses data from 685 PhD students in life sciences at Japanese universities, all of whom graduated between 2000 and 2011. The data includes dissertations, curricula vitae, and publication records up to 2018. The main variables include the degree of specialization during the PhD period, which is measured using the Herfindahl–Hirschman Index based on the concentration of research topics in published articles. Career outcomes are assessed in two ways: staying in academia after PhD (short-run) and obtaining tenure promotion (long-run). The study also controls for factors like student attributes, professor characteristics, and the organizational context of the universities.

Results: Short-term Effects (Staying in Academia): Specialization during the PhD period is found to positively influence the likelihood of staying in academia. The research suggests that students who specialize during their PhD are more likely to find an academic job due to the depth of expertise they develop, making them attractive candidates for entry-level academic positions. This aligns with the idea that specialization helps students build a solid knowledge base and academic profile quickly, making them competitive for early-career positions.

Long-term Effects (Tenure Promotion): The relationship between PhD specialization and tenure promotion is more complex, with an inverted U-shaped pattern. The analysis shows that moderate specialization during the PhD period increases the likelihood of obtaining tenure promotion, while extreme specialization (over-specialization) reduces the chances. This suggests that while a certain level of specialization is beneficial for establishing a solid academic foundation, excessive specialization may hinder the development of broader, more flexible skills needed for higher academic positions and independent research later in a career.

Furthermore, specialization in the years following the PhD is found to have a negative impact on tenure promotion. This suggests that as a researcher’s career progresses, specialization can limit the flexibility required for creativity and innovation, which are essential for advancing to senior academic roles.

Discussion: The paper emphasizes that the effects of specialization are contingent on the stage of a researcher’s career. During the PhD, specialization helps students gain expertise quickly, enhancing their early career prospects. However, as their careers advance, the need for broader, more flexible knowledge becomes critical, and over-specialization can hinder progress. The findings suggest that PhD students and academic institutions need to carefully balance specialization and generalization to optimize both short-term career success and long-term career development.

The study also underscores that the mechanisms of specialization’s effects differ across career stages. During the PhD, specialization supports skill acquisition and knowledge depth, which is essential for securing an academic job. In contrast, in later years, specialization can limit creative performance and innovation, potentially reducing the chances of achieving tenure.

Conclusion: This paper contributes to the literature on the organization of science and PhD training by providing insights into how specialization affects scientists' careers. It shows that while specialization is advantageous in the short-term (i.e., securing academic positions right after PhD), over-specialization in the long term can hinder tenure promotion due to reduced creativity and flexibility. The findings suggest that both PhD students and their supervisors should be mindful of the potential downsides of over-specialization, especially as researchers progress in their careers. Furthermore, this research has implications for the design of PhD training programs, encouraging a balance between developing deep expertise and fostering a broader, adaptable skill set for long-term career success.

ABSTRACT. 1. Introduction

Nowadays, academic scientists do not exist in a social vacuum or isolate themselves in the ivory tower. Over recent decades, the increasingly complex nature of modern science has encouraged academic researchers to collaborate with a wide range of individuals within and outside academia, especially in science, technology, engineering, and mathematics (STEM) fields (Bozeman et al., 2013). Recognizing this, governments and funding agencies worldwide, such as the U.S. National Science Foundation (NSF), have considered the broader societal impacts of research as an important science funding criterion since the late 1990s (NSB, 2011). Given the substantial public investment in science, there is an evolving expectation that STEM research should not only advance the scientific frontier but also bridge the gap between academia and society (Link & Wagner, 2021).

Indeed, science funding plays a crucial role in shaping today’s academic research landscape, stimulating collaborations that transcend traditional disciplinary and institutional boundaries (Arnott et al., 2020). As such, a variety of innovative research approaches have gained significant visibility in scientific practice, such as use-inspired basic research, ‘Mode 2’ science, and knowledge co-production (Howarth et al., 2022). Regardless of variations in terminologies, these approaches often involve collaboration between academics and non-academics to generate and/or translate scientific knowledge (Nguyen et al., 2020). These collaborative efforts are perceived to be effective means to align research with societal needs, thereby enhancing the use and impact of knowledge (Safford et al., 2017).

Despite the potential benefits, research collaborations with non-academics have not been widely acknowledged or institutionalized in today’s scientific enterprises (Doberneck, 2016; Hart & Silka, 2020). These efforts are often hindered by conventional disciplinary cultures and academic reward systems, which typically prioritize research excellence over the additional efforts required for external collaborations (Klein & Falk-Krzesinski, 2017). Some critics further view these collaborative practices as a departure from the traditional focus on pure scientific inquiry (Henkel, 2005). This ambivalence reflects the tension between the established academic norms and the growing expectation for academics to collaborate with non-academics. Faculty tenure and promotion, after all, usually hinges on publication-based evaluation criteria, which may not fully capture the value of research collaborations with non-academics.

Why should one care whether research collaboration with non-academics affects faculty tenure and promotion processes? One simple reason is that the benefits of such collaboration for individual scientific careers are more often assumed than empirically examined. Drawing from the scientific and technical human capital (STHC) theory, this study focuses on the impacts of these research collaborations on the career progression of individual STEM academic scientists, rather than on societal or economic benefits writ large. Using a national survey of STEM faculty in U.S. research institutions and their bibliometric data, this study employs event history analysis to address a key question:

During the course of academic scientists’ career life cycle, how does research collaboration with non-academics affect their career advancement?

Investigating this question will help disentangle the relationships between the two, shedding light on the potential benefits and challenges of such collaborations for STEM academic scientists across different career stages. Furthermore, findings will offer valuable insight for science and university policymakers, identifying practices that could improve scientific careers and, ultimately, the capacity of individual academics to produce new knowledge.

2. Data and Methods

The data for this study are drawn from two major sources: the 2012 NSF-funded national survey (NETWISE II) of U.S. academic scientists and engineers, and bibliometric data of the survey respondents. The NETWISE II survey targeted 9,925 faculty from 521 academic institutions in four STEM disciplines: biology, biochemistry, civil engineering, and mathematics. The sampling frame was stratified based on institutional type, faculty rank, and discipline, with an intentional oversampling of women. The survey was administered online and took around 40 minutes to complete, ending up with a total of 4,196 valid responses (response rate = 40.4%).

The NETWISE II study also gathered the publication histories of all survey participants. For this study, combining survey and bibliometric data creates a unique dataset that allows for in-depth analyses of publication patterns and important individual data. Given the significant variations in research priorities and faculty expectations across different types of institutions, this study specifically focuses on academic scientists employed at U.S. R1 and R2 research institutions. The final sample used in this study includes 1,237 academic scientists and engineers from 218 U.S. research institutions.

3. Findings and Conclusion

Overall, the event history analyses show that academic scientists who have ever collaborated with non-academics will achieve higher ranks to tenure and full professorship more quickly. Based on bibliometric data, these results provide evidence that knowledge-focused NA collaboration can benefit academic careers, though the effect is weaker for promotion to full professor. In other words, interestingly, the positive effects of NA collaborations on career progression are particularly pronounced for early-career academics. From the resource-based viewpoint, one possible explanation is that having these experiences provides opportunities for them to develop professional networks and build a diverse research portfolio that is beneficial to career development (Bozeman & Corley, 2004).

It should also be noted that the magnitude of NA collaborations plays a crucial role in academic career advancement. As expected, results indicate that collaborating more intensively with non-academics, rather than with academics, significantly slows down early-career scientists’ progression toward tenure. This can be attributed to the fact that collaborating with non-academics tends to be a time-consuming research practice involving uncertainties/risks that may delay publications, which is particularly an issue for early-career academics as they need to meet the traditional merit-based tenure criteria (Gulbrandsen & Thune, 2017; Pannell et al., 2019). Contrary to the initial expectation, NA collaboration intensity has no significant effect on promotion outcomes for senior academics, which is instead determined by scholarly productivity. In sum, these mixed findings demonstrate that it is important for academic scientists to maintain a balance between NA collaboration and traditional academic pursuits, especially during the early stages of their careers. If costs outweigh benefits in some cases, NA collaboration can hinder individual career progression, instead of accelerating the process.

ABSTRACT. Cultivating innovation and entrepreneurship (I&E) outputs from universities into society is critical in realizing the broader impacts of public and private investments in university research. Patents are an important output of this process, in which university research is commercialized into services and products that benefit society. Since the passage of the Bayh-Dole Act of 1980, which facilitates the patenting by U.S. universities of I&E derived from federally-funded research, university administrators have increased their efforts in producing patents. However, while it is clear university administration value I&E work, a pertinent question is, do the faculty members who make promotion and tenure decision equally value I&E work?

In a series of three studies, we examine how, when and why academic promotion systems disincentivize I&E, and what can be done to alleviate misalignment between I&E and what faculty value in promotion. In the first study, we examine whether faculty are rewarded for I&E activities in the promotion and tenure (P&T) process, the primary reward system in universities. Using data from 811 faculty members in STEM disciplines and 4875 ERLs submitted as part of P&T packets, we found that obtaining a patent or discussion of I&E activities in a candidate's external review letters (ERLs), was related to a larger percentage of unfavorable P&T outcomes, including a 68% reduction in the likelihood of a positive vote from the provost for candidates with patents.

In two follow up studies we examine the underlying beliefs driving these negative perceptions of I&E and how to improve them. The first was a survey of P&T decision makers which found support for a belief-based model. Most notably, we found perceptions of the value and importance of I&E was associated with the extent to which faculty believed social impact and applied research was a goal of science/scientists. In the next study, we examined intervention pathways to improve how faculty value I&E in promotion decision-making. We found that an easy-to-implement rubric intervention highlighting how I&E benefits society led faculty to value I&E more positively.

Together, these studies demonstrate the critical need to create a more inclusive academy that rewards a diversity of faculty outputs – including I&E work – for universities to succeed in generating societal and economic prosperity for the communities they serve. In our conclusion, we outline recommendations to address resistance to I&E and help realize the broader impact of science in society by aligning rewards for I&E activities with promotion systems.

ABSTRACT. Scientific peer review fundamentally shapes scholarly discourse and researchers' career trajectories, particularly for early career scholars.  Growing concerns around a "peer review crisis," including concerns around biases in scientific publishing and difficulties finding suitable reviewers, raise important questions about equity and efficiency in the peer review process. Leveraging large language models (LLMs) within an augmented resume audit design, we first experimentally test how explicit author identity cues influence LLM-generated editorial evaluations of identical scientific papers. Initial findings reveal some systematic biases: papers attributed to female authors, those affiliated with lower-prestige institutions, and—to a lesser extent—marginalized racial groups receive lower quality scores and a greater number of critical, yet less constructive, reviewer comments.  We further evaluate the quality of LLM-generated comments and assess feasibility in serving as a substitute or compliment to human-generated reviews.  LLM generated reviews are promising, albeit inconsistent. LLM biases may increase barriers to scientific productivity and career advancement, highlighting broader implications of LLM-based bias detection methodologies applicable to other areas of research.

ABSTRACT. How does experimentalist governance affect firms innovation? What drives the world's largest oil and gas firms to invest in Carbon Dioxide Removal (CDR) technologies?

Public Policy scholars have recently put a lot of thinking into CDR and its techno-economic attributes as drivers of corporate engagement and investments (Battersby et al., 2022; Beaumont et al., 2022}. However, the early role of regulators in shaping the nexus between firms and innovation in CDR technology in the oil and gas industry has been largely neglected. This paper argues that early experimentalist CDR-related policies in the U.S. have shaped strategic corporate decisions, particularly in the realm of innovation and intellectual property. We center our argument on the regulatory drivers behind this innovative behavior in the context of high risks, uncertain returns, and the absence of a strong market for CDR (Sabel & Victor, 2022).

By fitting an Interrupted Time Series (ITS) model on patent data spanning from 1975 to 2024, this study reveals a notable shift in innovation activities among fossil fuel firms following the enactment of 45Q. Specifically, results ITS suggest an immediate effect of approximately 98 more patents per year across 50 firms after the year of policy enactment until 2024. In addition, using ARIMA model I forecast the counterfactual scenario—what the trend in patent filings might have looked like if the 2008 intervention had not occurred. By fitting the ARIMA model using pre-2008 data, I generate a forecast that helps to isolate the effect of the intervention. The main results suggest that without targeted regulatory incentives, fossil fuel firms are unlikely to prioritize or invest significantly in climate mitigation technologies.

This argument aligns with recent discussions on state experimentalist interventions and technology change, suggesting that government-led policies, driven by state interests, have initiated processes of policy learning and reductions in technology costs (Victor & Carlton, 2023; Breetz et al.,2018; Allan et al., 2021; Meckling, 2021). In the case of CDR, early policies in the U.S. aimed to redirect the technological focus of the largest oil and gas firms towards innovations in CDR. Oil and gas firms responded to early policies by increasingly applying for patents, reflecting a strategic alignment with the evolving regulatory landscape and the anticipated future significance of CDR technologies. As firms might expect that their assets will be reevaluated as existential risks from climate change become more prominent (Colgan et al., 2021), some view oil and has firms’ interest towards CDR as a (1) “mitigation deterrence” (Carton, 2023) or (2) as a way to shift the business model from production to removal of carbon. For a "brown-to-green" transition to be successful, fossil fuel firms need to reorient their behavior toward decarbonization. State-led interventions must overcome entrenched market dynamics and strategic resistance from these firms.

In conclusion, this paper explores how early experimentalist policies, particularly the 45Q tax credit introduced in 2008, provided a unique regulatory push that incentivized U.S. fossil fuel firms to invest in CDR technologies. Despite facing initial challenges such as limited credit availability and uncertainty in allocations, 45Q emerged as a strategic tool for the government to address high-risk, capital-intensive solutions like CDR.

ABSTRACT. INTRODUCTION The transformation of traditional industries through innovation processes to make them cleaner and more sustainable is already visible in the science and policy narrative (NAS, 2024; Draghi, 2024) as a key step for a cleaner more sustainable World Our research focuses on the construction industry and it analyses the most innovative practices that develop clean cement. It understands the reasons of these practices and uncovers the innovation and public policy role in these uses in the last decades in Latin America. The research is based in the ongoing work of the Ibero-American Network for Carbon Neutrality in the Cementitious Materials Industry (RINCIMCI https://www.cyted.org/RINCIMCI), whose coordinator is one of the authors; and the CAPES Print scholar post-doctoral work on science and innovation policy, that the other author has undergone during the past year. The Ibero-American Network for Carbon Neutrality in the Cementitious Materials Industry (RINCIMCI) is carrying out groundbreaking research to understand the use of clean and innovative ways in the production of cement in the Ibero-American countries, i.e.: Argentina, Brazil, Chile, Colombia, Spain, Peru and Portugal, including 17 research teams in a dozen of research institutions and companies spread around the network. Cement production is responsible for 8% of global carbon dioxide (CO2) emissions, therefore the transformation of the cement industry plays a crucial role in reducing emissions to mitigate climate change (Scrivener, 2018). The optimal combination of low-carbon measures could contribute to reducing annual emissions by around 65% in 2050 and cumulative emissions by around 48% over 2020-2050. (Cheng et al. 2023)

METHODS First, we have reviewed from the Scopus database all the published literature between 2021-2024 related to the construction industry and use of Clinker, Cement, Concrete, Construction and (re)Carbonation (elements of the construction industry). We have identified and mapped who, what entities and in which countries are actively involved, with a specific focus on the industrial actors responsible for the cement production. We have selected the top countries more active in these practices. Then, we have reviewed the regulations (including national laws and strategies/programmes) that these countries have in place (whenever so) to guide the construction industry regarding the cement production. We have identified the themes of “cement production”, “construction industry”, “clean technologies”, “CO2 emissions” relevant for our research. Finally, we have carried out semi-structured realist interviews (Rees et al, 2024) to the big industry representatives participating in the RINCIMCI network to understand how (and if) the policy has had any effect (Dino, 2022) in using cleaner techniques during the period of study (2021-2023), and why is that. The interviews are carried in the different countries that show different policy context. The preliminary findings on the research are presented below. We aim to present an advanced study at the time of presentation

FINDINGS As a novelty, the findings from this research reveal how the construction industries in the Latin-American countries are introducing new practices towards lower carbon emissions and sustainability and how the public policies are effecting these practices. Cement contributes to 8% of overall human-made greenhouse gas emissions. Substituting clinker in cement with sustainable material can reduce CO2 emissions by 40%, saving up to 500 million tons of emissions by 2030 (WEF, 2023). More evidence on different materials is available (Rêgo, 2015). Our research brings out that the use of sustainable cement production, such as widely addressed in the RINCIMCI network, focuses its main outcomes in Brazil and Colombia. These are two countries with two different settings in the construction industry, and two different ways of tackling innovation policies for the period under study. Evidence from our research checks on the sustainability policy during 2021-2023 while the two countries changed their governments. The research uncovers the effects of the public policy in general, and innovation policy more specifically, in these practices and settings. At the present moment we are carrying out the qualitative analysis to understand the reasons behind this use by interviewing with the five industries involved in the cement production in the two countries. While RINCIMCI will last for three more years, this research aims to conclude in spring 2024.

REFERENCES Cheng, D., Reiner, D. M., Yang, F., Cui, C., Meng, J., Shan, Y., ... & Guan, D. (2023). Projecting future carbon emissions from cement production in developing countries. Nature Communications, 14(1), 8213. Dino, A. (2022). A Realistic Evaluation of Science Policy-Generating Learning for Spanish Public Administration Institutions (Doctoral dissertation, UCL (University College London)). Draghi, M. (2024). The Future of European Competitiveness Part A: A competitiveness strategy for Europe. Habert, G., Miller, S. A., John, V. M., Provis, J. L., Favier, A., Horvath, A., & Scrivener, K. L. (2020). Environmental impacts and decarbonization strategies in the cement and concrete industries. Nature Reviews Earth & Environment, 1(11), 559-573. National Academies of Sciences, Engineering, and Medicine. (2024). Carbon Utilization Infrastructure, Markets, and Research and Development: A Final Report. Rees, C. E., Davis, C., Nguyen, V. N., Proctor, D., & Mattick, K. L. (2024). A roadmap to realist interviews in health professions education research: Recommendations based on a critical analysis. Medical Education, 58(6), 697-712. Rêgo, J. H. S., Nepomuceno, A. A., Figueiredo, E. P., & Hasparyk, N. P. (2015). Microstructure of cement pastes with residual rice husk ash of low amorphous silica content. Construction and Building Materials, 80, 56-68. Schneider, M. (2019). The cement industry on the way to a low-carbon future. Cement and Concrete Research, 124, 105792. Uratani, J. M., & Griffiths, S. (2023). A forward looking perspective on the cement and concrete industry: Implications of growth and development in the Global South. Energy Research & Social Science, 97, 102972. World Economic Forum (2023) https://www.weforum.org/stories/2023/10/this-new-material-could-change-how-we-make-cement-forever-and-cut-500-million-tonnes-of-emissions-by-2030/

AUTHORS Dr Armela Dino is a Science Policy Analyst at the Spanish Ministry of Science, Innovation, and Universities (on leave) and a lecturer of International Economics at the University of Complutense de Madrid (UCM). Contact: adino@ucm.es https://sites.google.com/view/armeladino

Prof João Henrique da Silva is an Associate Professor of Civil Engineering at the Federal University of Brasília (UnB) and coordinator of the RINCIMCI international network. Contact: jhenriquerego@unb.br https://scholar.google.com.br/citations?user=JrhoznMAAAAJ&hl=pt-BR RINCIMCI https://www.cyted.org/RINCIMCI

ABSTRACT. The connection of distributed renewable energy sources to conventional power systems is a new task in practice for policymakers as well as grid managers, especially due to stochastic supply characteristics. This paper aims to understand the utility of contemporary artificial intelligence techniques in energy demand forecasting about renewable energy and grid robustness policy creation and implementation. The paper fills the existing literature gap by providing an investigation of the potential for implementing AI-supported forecasting models within the policy-based model operating in the context of the combined problem of regulatory compliance and sustainable development.

The study utilizes both a quantitative examination of high-temporal resolution energy use data from three large American cities and qualitative surveys of policy barriers. To this end, we designed and empirically evaluated a complex Deep Learning model capable of processing multiple data feeds: historical electrical usage, temperatures, socio-economic statistics, and the live status of the grid. The model outperforms statistical methods in terms of mean absolute percentage error and the peak demand forecast, where the results have been greatly enhanced.

Our methodology encompasses three key components: (1) The creation of a Temporal Pattern Recognition System that fuses LSTM networks with self-attention mechanisms of the transformer model; (2) The application of a policy-driven optimization algorithm that ensures policy compliance and breaks through sustainability barriers; (3) The construction of an explainable AI system that allows policymakers to understand the reasoning of the presented policies and actions.

The results show that the robotic systems in forecasting improve the capabilities of grid management and support the integration of renewable energy. The utilization of the proposed approach allowed obtaining improved predictions of renewable energy generation at various time horizons, which can be used for more efficient grid balancing. Some evident and quantifiable changes that originated through the application of the AI forecasting system are a decrease in grid stabilization costs and an increase in the efficiency of the use of renewable energy. The organisational policy-conscious design of the model made it possible to improve power distribution for consumers with respect to legal requirements on environmentalism and goals for sustainability. Generic surveys with the different stakeholders showed that trust in decisions regarding grid management improved when based on interpretable AI outputs, including the abilities of policymakers.

Our study also responds to a number of significant problems that are currently experiencing development in the research area: data quality measurement, natural language explanation for the interpretation of different models for policy use, and the necessity of quantifying the uncertainty of forecast results. The work provides fresh approaches to these difficulties, the data validation architecture that is tailored to challenges of the energy sector, and the method for the uncertainty specification that delivers the confidence intervals for the results of forecasting.

The implications of this research do not stop with the analysis of accurate technical advancements but engage with broader discussions. The results imply that AI-based energy demand forecasts can act as an important link between on the one hand the highly technical-oriented field of grid management and on the other the goals and visions of policymakers. This work gives a clear explanation of how the hosts of the AI models at an advanced level can be incorporated in the making of policy while keeping in mind the uncertainties of the general public. Resulting from this integration has been enhanced decision-making concerning the prospects of renewable energy integration and the systems of support for grid resilience.

One innovative aspect of this work is the identification of a policy-based approach for evaluation of AI forecasting models where such models are not only evaluated on accuracy and data but also from the standpoint of how they would be useful in achieving policy objectives and fulfilling regulatory obligations. This framework enables assessment of the success of AI systems in attaining the technical as well as policy objectives set in the power industry.

The study also explores the general implications to energy policy advancement, where the potential of AI-analysis is explored for its impact on the general policies, regulatory rules, and sustainability endeavours in the energy market. This paper finds that the implementation of AI tools for forecasting creates more adaptable policies to shifting energy environments while maintaining the stability of utility and power distribution.

The future research direction suggested in this study is the form of cross-sectional cooperation required to manage the grid by integrating different jurisdictions; the second is about the need to standardize the policy AI system metrics; and the third is about the new form of smart grid challenges such as cybersecurity and data privacy.

This paper deploys the existing literature on the use of smart technologies in the formulation of energy policies and grid management to present practical lessons to policymakers and grid operators. The study also shows that AI-based forecasting systems are highly capable of transforming policies in the energy sector and their applications for integrated renewable power systems and grid reliability improvement.

ABSTRACT. Climate change and environmental degradation are making increasingly urgent a green transition towards an economy where human-produced greenhouse gas emissions are reduced to (net) zero, and natural resources are used more efficiently. The green transition has become a prominent focus of policy across the globe and particularly in Europe, where it has been recently associated with the digital transformation. Starting from the new EU Industrial Strategy (European Commission, 2020), and moving towards the new Green-Deal Industrial plan (European Commission, 2023), the “green and the digital challenges” have been claimed to merge into a “twin transition”, in which digitalisation serves to enhance environmental sustainability (JRC, 2022).

However, amidst the recent surge in policy recommendations advocating for this twin transition, concerns regarding its territorial distribution across regions have arisen. On the one hand, regions are shown to have different greening capacity (e.g. Barbieri et al., 2020) and to face different direct (e.g. phasing out of brown energy) and indirect (e.g. reallocating labour from brown to green occupations) costs in moving along the green transition (Rodrıguez-Pose & Bartalucci, 2023). On the other hand, digitalisation also unfold irregularly across places, revealing regional gaps in the development, adoption and exploitation of new digital technologies (e.g. Corradini & De Propris, 2017). In light of that, the combination of the two transitions can also be expected to differentiate across regions, rendering the investigation of its geography of paramount importance.

Research on the geography of the twin transition is still thin, but fast growing (e.g. Faggian et al., 2014). In geography of innovation studies, the analysis on the twin transition phenomenon is carried out mainly to investigate the extent to which new digital and/or green technologies emerge from pre-existing regional technologies (e.g. Bachtrogler-Unger et al., 2023). In this paper, building on the recombinant theory of innovation, we delve into the nature and patterns of combining green and digital technologies, specifically examining the speed at which regions integrate these technologies. More precisely, we are interested in analysing how fast regions are capable to implement combinations of green and digital knowledge in their local inventions, compared to what happens at the frontier of the knowledge space. This aspect of the twin transition has not yet been addressed in literature. However, we consider it highly relevant given the urgent pressure from policymakers for regions to swiftly advance their combined green and digital transformations.

In our analysis, we examine two key aspects. First, we analyse how the speed of combining green and digital technologies is influenced by the characteristics of the relative regional knowledge, as well as by the nature of the combinatory patterns of the two kinds of technologies. This will be useful in identifying potential leverages that policy-makers can exploit to speed up the twin transition. Second, we investigate how the combination pace of green and digital technologies affects the impact of local twin inventions on the development of subsequent technologies. This aspect is also important to disentangle if the regions’ twin speed can affect their capacity of serving as reference for the development of new technological paradigms. Using the European Patent Office (EPO) PATSTAT dataset, we build on previous patent-based research to identify green and digital technologies (Barbieri, 2016; Martinelli et al., 2021). We then define twin (green and digital) inventions based on the co-occurrence of digital and green (5-digit CPC) classes in patents (Kogler et al., 2013). These twin inventions are mapped across NUTS3 regions in 28 EU countries, covering the period from 1999 to 2019. We propose a novel measure of the average speed at which a region closes its gap in combining green and digital technologies relative to the frontier, that is, with respect to when the combination first takes place.

We find that the regional speed of combining green and digital technologies increases when the region has experience in dealing with diverse and complex twin-transition-related knowledge. Furthermore, we find that when the green and digital domains share common interdependencies, the speed of combination is higher. Third, the technological focus of the inventive teams, rather than their size, leads to quicker combinations. We discuss the counterintuitive evidence according to which faster combinations occur in case the technological distance is larger. This effect is explained by high density structures in the knowledge network, captured by a transitivity coefficient: intuitively, the longer the recombination path, the more are the opportunities to find short cuts that speed up twin combinations. As far as the impact of fast twin combination is concerned, we find that this is connected to a larger number of forward citations, only for granted patents.

References

References Bachtrogler-Unger, J., Balland, P.-A., Boschma, R., & Schwab, T. (2023). Technological capabilities and the twin transition in Europe: Opportunities for regional collaboration and economic cohesion. Technical Report 117679, MPRA Paper

Barbieri, N. (2016). Fuel prices and the invention crowding out effect: Releasing the automotive industry from its dependence on fossil fuel. Technological Forecasting and Social Change, 111, 222–234

Barbieri, N., Perruchas, F., & Consoli, D. (2020b). Specialization, diversification, and environmental technology life cycle. Economic Geography, 96(2), 161–186

Corradini, C. & De Propris, L. (2017). Beyond local search: Bridging platforms and intesectoral technological integration. Research Policy, 46(1), 196–206.

European Commission (2020). A new industrial strategy for Europe

European Commission (2023). The European green deal industrial plan

Faggian, A., Marzucchi, A., & Montresor, S. (2024). Regions facing the ‘twin transition’: combining regional green and digital innovations. Regional Studies, 1-7.

Kogler, D. F., Rigby, D. L., & Tucker, I. (2013). Mapping knowledge space and technological relatedness in US cities. European Planning Studies, 21(9), 1374–1391 JRC (2022). Towards a green & digital future. Joint Research Centre (JRC), European Commission.

Martinelli, A., Mina, A., & Moggi, M. (2021). The enabling technologies of industry 4.0: examining the seeds of the fourth industrial revolution. Industrial and Corporate Change, 30(1), 161–188

Rodríguez-Pose, A., & Bartalucci, F. (2024). The green transition and its potential territorial discontents. Cambridge Journal of Regions, Economy and Society, 17(2), 339-358.

ABSTRACT. One of the challenges that the National Science Foundation (NSF) seeks to address through multiple funding initiatives is broadening participation in the innovation ecosystem. However, supporting and growing the innovation ecosystem requires capacity-building of “innovation nodes.” We define “innovation nodes” as any entity which both contributes to and benefits from participation in the innovation ecosystem, especially in areas related to advanced, emerging technologies, such as automated vehicles. We are particularly interested in the context of advanced automotive workforce development in Arizona.

Systematic thinking about the environmental context of innovation is lacking. A fictional scenario is often cited where appropriations in the House Science and Technology Policy Committee are made with the underlying assumption that adding more funds to a particular program will directly entail a significant innovation. The problem, however, is that, with the complexity of the evolving NIS, contexts play a major role in shaping the system. Therefore, strategic policy making will be enabled by deeper understanding of contexts, connectivity and interactions.

The major research questions that we seek to address are:

1. What are the major operation categories of innovation nodes? 2. What are the contextual characteristics of each node? 3. How do various node types respond to public policy initiatives?

In search for ways to create a shared understanding of the impact of context on innovation, we seek to build a prototype of a visualization tool that will (1) map the landscape of interrelated innovation nodes; (2) iteratively analyze and develop narratives of potential changes in the interrelations between system elements; and (3) engage different stakeholders in the co-creation and exploration of solutions. The dramatic enhancement of interactive visualization technology allows the organization of complex data to explore solutions and answering “What If” questions. Most importantly, interactive visualization technology can bring together different stakeholders, such as researchers, policymakers and the business community, to work together in transdisciplinary partnership settings. Ultimately, this project seeks to develop a prototype of a visualization tool which is crucial in organizing different stakeholders to better understand the regional innovation ecosystem.

However, a pre-work is needed to establish this tool. The research method utilizes a convergence framework in two phases. Phase 1 will seek to collect qualitative data from the local community, which will feed into the visualization tool; and Phase 2 will seek to develop the interactive visualization tool. The tool will be modifiable for other regions beyond the locality of the State of Arizona. In this conference, we report the findings from Phase 1 only.

Phase 1 involved qualitative data from key stakeholder using the conceptual framework of Lean Launchpad, originated by Osterwalder & Pigneur in 2010. This framework enables the development of a language, model and shared vision to better understand change levers within the innovation ecosystem.

Participants were recruited from organizations of different orientations: public and private, those who are at IHE and those who are leading the advanced automotive industry in Arizona. Qualitative data were collected through a structured workshop that integrated focus group discussions among various stakeholders in rotating sessions. The workshop took place with support from Arizona Commerce Authority (ACA), Science Foundation Arizona, and the Institute of Automated Mobility (IAM) in Arizona, “the first state to test and deploy self-driving cars on public roads.” Science Foundation Arizona is a non-profit organization staffed by ACA employees with mission-driven research. The Foundation is a public-private organization dedicated to diversifying and strengthening Arizona’s economy. Both the Foundation and ACA have formed a partnership to accelerate industry-education collaboration and workforce development—an important context for this study.

To collect data for Phase 1, the workshop proceeded with an embedded research goal to answer the research questions. Our effort in designing the one-day workshop was to transcend barriers to create a better understand of the system by recognizing the complex influences on changing contexts. We framed the mission at the beginning of the workshop as a response to the need to develop a coordinated curricula at Arizona universities and community colleges that support that advanced automotive industry. The industry broadly includes Connected and Automated Vehicles (CAVs) and Zero-Emission Vehicles (ZEVs).

To start the discussion, moderators identified the fast-growing population and excellent IHE as drivers for the State’s leading edge in workforce development. Arizona, for example, holds the fourth highest growth rate in education attainment and the fifth highest growth in skilled jobs, with an average of 260 people moving to Arizona every day according to a recent Census study. In 2023, Arizona’s automotive industry was first in jobs added since 2019 (135% growth) and second in average salary ($86,008). Many Original Equipment Manufacturers (OEM) have Proving Grounds in Arizona. In addition to year-round operation potential with minimal inclement weather, supporting regulations and law enforcement interactions, the State is positioned to be a leader in workforce development in the sector.

After establishing the potential of the effort, the focus groups were moderated with semi-structured interviews. There are three major themes that emerged, which are supported in the literature in other areas (e.g., the semiconductor industry). First, the need for an organization to initiate and coordinate the effort of workforce development; neither industry alone nor IHE alone can establish work at the interfaces for various reasons. Second, while content expertise is available, both proprietary issues and lab access remain a challenge in training the workforce. Third, there was a recognition of the need for upskilling: on the end of current technicians (to adapt to the new technology requirements) as well as on the faculty side (to spend sabbatical time in industry to update knowledge and pedagogies).

In summary, developing an understanding of the innovation ecosystems tends to be based on demarcation between the potential and contributions of the government, the various levels of industry partners (big and small), and the various levels of output of the IHEs. While addressing these problems is more difficult than it seems, having a visualization tool of the innovation ecosystem, and the interaction between its nodes, is one significant step toward understanding the remarkably complex system.

ABSTRACT. 1. Background and Motivation

Entrepreneurial Ecosystems (EE) have gained traction over the last decade as an efficient framework for understanding the significance of entrepreneurial events beyond the focus on the individual, generating relevant insights how regional factors can affect the incidence of entrepreneurship in a specific locale (Brown & Mason, 2017)). This approach is based on the argument that entrepreneurship takes place as an intrinsically systemic phenomenon (Cao & Shi, 2021). Traditional EE appraisals have mainly dealt with individual ecosystems’ elements, often neglecting the complexity upon which these structures emerge (Roundy et al., 2018). Accordingly, the substantial literature on entrepreneurial ecosystems (EE) has yet to reach consensus on defining the exact nature of entrepreneurial ecosystems and their most critical components, limiting prescriptive analyses and incorporation of its insights for regional policy.

In recent years, literature has underscored these issues, recognizing that there may be multiple configurations of self-sustaining regional entrepreneurial ecosystems, thus allowing a better comprehension of EE heterogeneous configurations and evolutionary trajectories (Cherubini Alves et al., 2021). However, as research moves forward in this respect, myriad configurations emerge in different parts of the world (Cherubini Alves et al., 2021; Fischer et al., 2024; Komlósi et al., 2024; Schrijvers et al., 2024), generating a sense of disconnection between EE conceptual landmarks and the corresponding body of empirical findings. So far, no systematic effort has been undertaken to generate robust alignment in this discussion.

Our goal in this research is to derive a typology of entrepreneurial ecosystem configurations across U.S. metropolitan and micropolitan regions and, importantly, how they change over time. By engaging in this comprehensive and longitudinal exercise we should be able to address the following research questions: (i) What are the predominant features of EE configurations? (ii) How do EEs change over time, and what patterns emerge from that evolution? Ultimately, this should lead to consistent method for detecting, classifying, and assessing the trajectories of EEs across alternative configurations.

2. Method

To address the issue of entrepreneurial ecosystems configurations, we draw from widely accepted conceptualizations of entrepreneurial ecosystems (e.g. Stam & van de Ven, 2021) and combine them with “bottom-up” data analysis. Our approach seeks to understand the levels of necessity and sufficiency in ecosystems’ elements, which elements can substitute for each other, and what elements are complimentary.

Many of the most advanced ecosystems in the US are densely resourced in all nine categories, which we refer to as ‘canonically complete’ ecosystems. However, many of the other ecosystems are incomplete, but have the potential to develop or attract missing elements. Other ecosystems may have just a few elements that combine to provide unique opportunities. We will use this framework to build a dataset of relevant proxy measures of EE dimensions gathered from multiple data sources. We will use the 381 US metropolitan statistical areas (MSAs) based on economic activity to identify potential entrepreneurial ecosystems. We will then test broadly available indicators as potential outcome variables, such as the rate of start-up formation or the growth in SBIR awards to each region, to find indicators that distinguish EE configurational trajectories.

We will use fuzzy-set qualitative comparative analysis (fsQCA) to delineate the different configurations of entrepreneurial ecosystems. QCA allows the identification of configurations or ‘recipes’ of causal conditions associated with different outcomes, following the equifinality principle, i.e., multiple combinations that lead to similar outcomes. The ‘fuzzy set’ allows for the inclusion of quantitative and qualitative non-binary variables. For such metrics, calibration values or levels are defined by the researcher based on an interplay between theoretically-grounded criteria and statistical structure of data.

3. Envisaged Results

Expected findings will suggest alternative ways to nurture thriving ecosystems and support locations that are still seeking to gain traction. For instance, in the analyzed high-tech corridor, Fischer et al. (2021) found a first configuration entirely reliant on the availability of a competitive knowledge infrastructure (represented by research university campus, innovation habitats, and presence of knowledge-intensive jobs), while a second key configuration combined elements from the knowledge context (innovation habitats and Human Development Index) with characteristics of the socioeconomic system (investment attractiveness in knowledge-intensive industries). Infrastructure, human capital, credit availability and population size (approximating agglomeration economies) made up a secondary layer of important attributes. This analysis will enable decision-makers to characterize and measure the critical conditions and determine those necessary to ecosystem emergence across all configurations and those specific to certain configuration(s).

By disentangling the reasons why heterogeneous configurations co-exist and evolve among entrepreneurial ecosystems, we will be able to have a much clearer picture of what kinds of initiatives can be functional to enable entrepreneurial activity and innovation in places. Hence, this proposal comprehends state-of-the-art research that will allow taking a significant step forward in our knowledge concerning the configurational trajectories of entrepreneurial ecosystems. Based on findings from other contexts, we expect that analysis will show a broader range of ecosystem types and development paths. Plus, by applying fsQCA longitudinally in a variety of regional settings, we expect to derive a consistent set of configurations that can inform scholars and policymakers beyond the US context.

ABSTRACT. Recent waves of technological change are reshuffling not only the occupational and industrial composition of countries and regions, but also the nature of skills required by labour markets, in ways that are often hard to predict (Ciarli et al., 2021). At the same time, the type and diversity of skills available to firms are a major factor in explaining productivity gaps between top and median performers (Criscuolo et al., 2021). The changing nature of in-demand skills is strictly linked to the unfolding of technological trajectories: firms adapt their skill requirements in such a way as to fully exploit the existing technological opportunities, improving their market position. It is therefore essential to harness the potential of identifying emerging skills demanded by labour markets, and to identify which are the skills combinations that are most conductive to improved economic performance.

Against this backdrop, we link the emergence of new in-demand skills portfolio to the growth of productivit y in the UK and EU regions through a three-step strategy, relying on economic complexity methods. First, we identify the emergence of new skills across UK and European regions, for each region-industry occurence. To identify skills, we exploit data on more than 200 million Lightcast Online Job Advertisements (OJA) across 11 countries (Austria, Belgium, Denmark, France, Germany, Luxemburg, Netherlands, Norway, Sweden, Switzerland, UK) between 2014 and 2020. The OJA data includes information on occupations, skills, regional location and contract type. Occupations and skills are mapped to the ESCO framework (European Skills, Competences, and Occupations). We analyze anually the skill demand trajectories within region-industry, from broadest to most granular levels (NUTS3 and 4-digit ISIC levels, respectively). Considering a timeframe of seven years allows to identify longer time-trends of emerging skill specialisations. Our data shows especially well the dynamics in service sector industries, which increasingly changed during the considered time period, due to increased digital technology development and adoption (Ciarli et al., 2021).

Second, we use different combinations of in-demand skills, such as IT-skills, to identify similar regions in terms of their changing skills demand trajectory. This is achieved by building a bipartite network connecting industries in each EU region to skills. Instead of relying on the Revealed Comparative Advantage (RCA), we follow Bruno et al. (2023) in constructing a Bipartite Weighted Configuration Model to statistically validate the network links. The resulting network infers similarity between skills (based on co-occurrence in regions) and between regions (based on shared skills specialisation). It clearly captures the variety of the European industrial landscape. For instance, we expect that Dutch regions, such as North Holland, demonstrate a strong specialization in IT and logistics-related skills, driven by their central role in trade and digital infrastructure.

From the validated network we obtain indices that quantify the internal coherence and diversity of the skills demanded by each region-industry. More precisely, we measure the connectedness of skills portfolios at the regional level, using information on the strength of the connection between skills obtained from the skills space. Moreover, drawing upon the economic complexity toolbox (Hidalgo & Hausmann, 2009; Tacchella et al., 2012) we construct a measure of skill-level sophistication. We expect that complex skills are hard to find, and that they occur in industries and occupations that require a wide set of skills.

Third, we link skills portfolios – characterised by their coherence and complexity – to the performance of industries across regions in terms of labour productivity. Data on labour productivity is available for NUTS3 regions from the Annual Regional Database of the European Commission's Directorate General for Regional and Urban Policy (ARDECO). The dataset is curated and maintained by the Commission's Joint Research Centre (JRC), based on information provided by the Eurostat's Structural Business Survey, and complemented by national and international data sources. By adding the productivity data to the skill-emergence analysis, along with controls for occupational and industrial structures as well as relevant socio-economic factors, we identify in-demand skills composition that are more associated with the productivity of EU regional industries. For instance, we expect that regions with a strong emphasis on digital and IT-related skills, such as those found in tech-driven hubs across Europe, consistently show higher productivity levels, highlighting the critical role of advanced skill portfolios in driving regional industrial performance.

In order to address potential endogeneity between the skill profile of regions and their economic performance, we exploit aggregate changes that are exogenous to individual regions and common across the sample, relying on a shift-share / leave-one-out instrument design. One such source of exogenous change is given by technological shocks, such as the uptake of ground-breaking and pervasive general-purpose technologies, like artificial intelligence (AI).

On the whole, we elaborate a differentiation and explaination of region-industry productivity through temporal dynamics of complex skill compositions. Our findings are expected to confirm the relevance of high skills (Criscuolo et al., 2021) and IT skills (Brynjolfsson et al., 2024) for labour productivity. Moreover, we expect new complex skills such as “machine learning” to appear in region-industry settings that require a high variety of related skills. With this new approach, we not only identify but also provide a new perspective on regional productivity, in the time of unprecedented, quick, and wide adoption of digital technologies.

ABSTRACT. Generative artificial intelligence (GenAI) has demonstrated remarkable capabilities in generating text, images, and other forms of content by leveraging large language models. Its potential applications span diverse domains, including automated decision-making, content creation, and personalized services. The release of OpenAI’s ChatGPT-3 in 2022 marked a pivotal moment, accelerating the adoption of generative models in commercial settings and highlighting their potential to revolutionize industries.

Despite its immense potential, GenAI also introduces risks that stem from its technical architecture and deployment environment. GenAI models operate as “black-box” systems, where their high-dimensional and complex neural networks make it challenging to interpret, explain, or correct outputs. This opacity is particularly problematic when GenAI is integrated into decision-making processes in critical sectors like finance and healthcare, where errors or biases could have serious consequences. In addition, the growing interdependence between developers and deployers, often through third-party collaborations, compounds the risks of data breaches, regulatory non-compliance, and cybersecurity vulnerabilities. It is crucial to understand what risks are involved in developing and deploying GenAI in different sectors and what approaches can be taken to properly govern the risks in facilitating innovation.

This study examines the implementation of GenAI in the financial services and healthcare sectors, focusing on adoption patterns, emerging risks, and risk management approaches. Using a sectorial innovation system framework, we analyze how different industries integrate GenAI into their decision-making processes and manage associated risks. A comprehensive approach is taken to data collection by conducting 37 interviews between February and May 2024 with stakeholders across financial services, healthcare, technology firms, and regulatory bodies. A combination of stratified, purposive, and snowball sampling is adopted to ensure representative data collection across sectors.

The financial services and healthcare sectors provide a compelling basis for comparative analysis due to their distinct regulatory environments, priorities, and systemic risks. Financial institutions prioritize maintaining market stability, consumer trust, and regulatory compliance, while healthcare institutions focus on patient safety, clinical accuracy, and data privacy. These differences influence how each sector adopts and manages GenAI technologies.

Our findings reveal distinct patterns in GenAI ecosystem development across sectors. In financial services, GenAI adoption has been driven by the need to enhance operational efficiency and decision-making. Use cases such as automated credit underwriting, personalized investment recommendations, and trading algorithms illustrate the sector’s reliance on generative models. However, explainability and performance remain central concerns. Financial institutions must ensure that decisions made by GenAI systems can be substantiated to regulators, clients, and other stakeholders, particularly to avoid legal liabilities under consumer protection laws. Moreover, the performance of GenAI models in trading and investment decisions has been scrutinized, as these systems are trained on historical data and may struggle to adapt to dynamic market conditions.

In healthcare, GenAI holds promise for applications such as medical diagnosis, research, and clinical support. However, the sector’s emphasis on patient-centricity introduces unique challenges. Healthcare professionals prioritize accuracy and robustness in GenAI outputs, as even minor inaccuracies can have profound consequences for patient outcomes. In addition, IT infrastructure limitations, including data privacy concerns and compatibility issues with electronic health record systems, hinder the scalability of GenAI solutions. The sector’s reliance on sensitive patient data also necessitates rigorous cybersecurity measures, further complicating the deployment of generative models.

This study identified two broad categories of risks associated with GenAI adoption. Technical risks include concerns about model accuracy, explainability, and robustness. For example, in financial services, explainability is critical for ensuring that generative models’ outputs align with legal and ethical standards. In healthcare, accuracy and robustness are paramount, as clinicians rely on GenAI systems to support decisions that directly impact patient care.

Implementation risks, on the other hand, stem from regulatory uncertainty, data protection requirements, and infrastructure constraints. In financial services, regulatory uncertainty is a significant barrier to scaling GenAI solutions. Institutions face difficulties in navigating evolving regulatory landscapes, which could render their investments in GenAI non-compliant. Outsourcing and third-party collaborations, while offering cost-effective solutions, exacerbate these challenges by introducing additional layers of risk related to data security and model transparency. In healthcare, infrastructure limitations and data-sharing restrictions hinder the broader deployment of GenAI. Hospitals often lack the necessary IT infrastructure and machine learning expertise to develop and deploy generative models in-house, leading to a reliance on partnerships with academic institutions.

This study explored two primary approaches to managing the risks associated with GenAI. Rule-based governance relies on formal regulations and standards to enforce compliance and accountability. For instance, financial institutions are subject to consumer protection laws and regulations that govern data privacy and cybersecurity. However, rule-based approaches can be rigid and may not adequately address emerging risks introduced by new technologies. In healthcare, strict data protection regulations govern how patient data is managed, adding layers of complexity to the adoption of GenAI solutions.

In contrast, principle-based governance emphasizes ethical guidelines and frameworks to promote responsible AI development. Approaches such as the OECD AI Principles advocate for transparency, trustworthiness, and human-centered design. While principle-based governance allows for flexibility and adaptability, it often lacks accountability mechanisms, making it challenging to enforce compliance. A mixed approach, combining rule-based and principle-based mechanisms, is essential for managing the risks of GenAI while fostering innovation.

The findings of our study reveal both similarities and differences in how the financial services and healthcare sectors approach GenAI adoption and risk management. Both sectors face challenges related to technical uncertainties and regulatory compliance, but their priorities and strategies differ. Financial institutions emphasize explainability and market stability, while healthcare organizations prioritize accuracy, robustness, and patient safety. These differences underscore the importance of sector-specific policy interventions.

Policymakers can draw valuable insights from this study to enhance GenAI governance. For financial services, regulatory frameworks should be adapted to address the unique risks posed by generative models, such as biases in credit underwriting and algorithmic decision-making. For healthcare, investments in IT infrastructure and data-sharing mechanisms can support the scalability and robustness of GenAI solutions. In both sectors, fostering collaboration between regulators, industry stakeholders, and academic institutions can facilitate the development of targeted governance mechanisms that balance innovation with risk management.

ABSTRACT. Previous work empirically demonstrated that the adoption of blockchain technology in Internet Courts significantly enhances worker efficiency, as evidenced by a reduced trial duration per case (Polzer et al., 2024). Building on these findings, this follow-up study investigates the broader and deeper impacts of blockchain and electronic evidence systems on court workflows, routines, and the skill acquisition necessary to integrate these technologies within the public sector. As unique institutions designed specifically for online legal disputes, China’s Internet Courts represent a pioneering model in digital justice, integrating advanced tools such as asynchronous hearing platforms and blockchain-based evidence management. Established to address the rising volume of digital and internet-related cases, these courts were the first globally to utilize digital technologies as primary tools in legal procedures, ensuring that cases are processed entirely online—from filing and evidence submission to hearings and verdicts. This innovative approach not only reflects the growing reliance on technology within the judiciary but also highlights the adaptability of public institutions to the demands of a digital society. Using in-depth interviews with judges and clerks from Guangzhou, Hangzhou, and Beijing Internet Courts, this study explores the human side of this efficiency-improved process: how adopting blockchain impacts day-to-day workflows, routines, and the learning curve for staff. The findings reveal a nuanced view of efficiency improvements: while blockchain has streamlined evidence handling and reduced trial duration, the integration of these systems affects other facets of work. Staff members report both enhanced workflow precision and increased dependence on digital verification processes, necessitating ongoing training and adaptation. Theoretically, this research draws on public administration concepts, particularly Lipsky’s (1980(Lipsky, 2010) street-level bureaucracy, to explore how technology affects the discretion and decision-making power of judicial workers. Additionally, the Technology Acceptance Model (Davis, 1989) and Job Demands-Resources (JD-R) theory (Bakker & Demerouti, 2007) frame the adaptation processes and psychological impacts of adopting these advanced systems. While TAM addresses how digital tools influence perceived ease of use and usefulness among court staff, JD-R theory provides insight into how the demand for continual learning and skill-building may intensify job demands, impacting the well-being of judicial workers over time. This study contributes to the literature on public administration and organizational behavior by offering a comprehensive analysis of the double-edged nature of technological innovation within public institutions. It highlights both the operational advantages and hidden costs of digital transformation, emphasizing the importance of managing technological transitions strategically to ensure they support the well-being of public servants while maximizing institutional benefits.

reference Bakker, A. B., & Demerouti, E. (2007). The job demands‐resources model: State of the art. Journal of managerial psychology, 22(3), 309-328. Davis, F. D. (1989). Perceived Usefulness, Perceived Ease of Use and User Acceptance of Information Technology. MIS Quarterly. Lipsky, M. (2010). Street-level bureaucracy: Dilemmas of the individual in public service. Russell Sage Foundation. Polzer, T., Steiner, R., Wu, N. J., Gasco Hernandez, M., & Buyannemekh, B. (2024). Digital Transformation in and of Local Governments – Variations on a Global Theme. Academy of Management Proceedings, 2024(1), 13604. https://doi.org/10.5465/AMPROC.2024.13604symposium

ABSTRACT. The advent of the internet has ushered in a new era of economic development and social transformation. At a national level, the Internet, particularly the emergence of the social media, has proven to increase GDP with such effects varying between economies (Czernich et al., 2011; Gruber et al., 2014), and to shape formal and informal institutions (La Ferrara, 2016), etc. At the firm level, the internet helps to improve business performance and employee productivity (Mack & Faggian, 2013). At an individual/household level, broadband has been found to improve both individual and family income (Ariansyah, 2018; Atasoy, 2013). However, another strand of literature has spotted the potential welfare losses it might bring to society (Goldfarb & Tucker, 2019), including political polarization (Levy, 2021) and the exacerbated inequalities.

Existing research highlights the “skill bias" created by the broadband internet adoption, where it complements skilled workers while substituting for low-skilled workers, thereby widening the wage gap (Akerman et al., 2015). Nonetheless, technological changes do not necessarily always favor skilled workers. Acemoglu (2002) posits that the impact of technology on the labor market depends on its relative profitability in replacing different skill groups. If replacing skilled workers with new technology is more profitable for the market, new technologies will attempt to replace skilled workers as well. A limited number of studies have investigated the effects of expanding the internet on inequality of income due to potential endogeneity issues. For example, the relationship between the Internet and social inequality is often bi-directional. On the one hand, national development and socioeconomic status greatly influence internet access (DiMaggio et al., 2001). On the other hand, there are mixed findings on the role of the Internet in reducing inequality (Chen & Wellman, 2005). Furthermore, there are numerous factors that are affecting both internet development and income inequality (the existence of confounders).

Despite the growing body of literature on the internet's impacts from various perspectives, evidence regarding its influence on intra-city inequality remains limited. One close study in recent years shows that higher internet penetration rates increase consumption inequality in counties across China in 2010 - 2016, but such effects are reduced when the internet penetration rate exceeds a certain threshold. Additionally, higher education levels help to offset some of these effects (Zhang et al., 2020). Current discussions on the relationship between the internet expansion and labor income underscore the role of technological advancements in exacerbating wage differentials among workers (Acemoglu & Autor, 2011; Acemoglu & Restrepo, 2019; Kogan et al., 2023; Kogan et al., 2020). Consequently, it is likely that the internet also plays a significant role in shaping the evolving patterns of income inequality, as possibly, wealth inequality in China.

This study seeks to address the existing research gap by offering the first empirical evidence on the impact of enhanced internet infrastructure on income distribution inequality in a developing context. We use data from China Family Panel Studies (CFPS) and China City Statistical Yearbook amid the staggered rollout of the Broadband China Strategic Program from 2014 to 2016, which is a natural experiment characterized by robust government support for the upgrade of regional internet infrastructure. Our findings reveal that the enhancement of internet infrastructure led to a statistically significant reduction in intra-city income inequality, as measured by the Gini index, by 0.035 points (control mean: 0.459, standard deviation: 0.061). This reduction is primarily driven by the program's impact on cities in central China. Furthermore, we observe that improved internet infrastructure increases household net income and labor income. Notably, the reason for the simultaneous increase in household income and reduction in city-level income inequality is that the income gains are concentrated among low- and middle-income households: The increase in income for the high-income families is significantly and substantially smaller. Therefore, the income gap is effectively narrowed. We also find that the expansion of broadband internet access reduced income inequality through its impact on the labor market. Specifically, the improvement of internet infrastructure facilitates the development of the platform economy, which creates new opportunities for low-skilled workers in sectors such as logistics and (online) retail.

This study contributes to the existing literature in several ways. First, it addresses a critical gap in understanding the relationship between internet access and inequality by examining the impact of enhanced internet infrastructure on intra-city income inequality. This paper uses Broadband China Strategic Program as a natural experiment to overcome endogeneity issues and provides novel evidence of its potential to reduce income disparities within cities. It further highlights the importance of considering the distributional effects of internet access within a developing context. Second, by focusing on the Broadband China Strategic Program, this study offers valuable insights for policymakers seeking to leverage technological investments to promote inclusive growth. Our findings suggest that cities’ efforts, such as targeted investments in broadband infrastructure, and the extra support from upper level governments can effectively reduce income inequality. This evidence supports the idea that digital inclusion policies can play a crucial role in ensuring that the benefits of technological progress are shared more equitably across society. Last, our analysis sheds light on the potential mechanisms through which better internet infrastructure can influence income distribution. By facilitating the growth of the digital economy and creating new employment opportunities for low-skilled workers, the upgraded internet infrastructure can empower historically disadvantaged groups and contribute to a more inclusive labor market.

ABSTRACT. Technological Forecasting and Social Change (TFSC) has been a cohort of submissions in science, technology, and innovation policy for decades. Covering a broad scope of interests in technological forecasting and modelling, management of technology, impact and evaluation of technology, and emerging technology & public policy, TFSC has established its distinctive research stands on methodology and practice of technology-oriented future studies and their interplays with social, political, and environmental factors. Clearly, the ATLC community is among TFSC’s key contributors, and we value this mature, comprehensive, and sustainable relationship.

This workshop aims to support scholars, particularly the ATLC community, in developing their research ideas and manuscripts, with the goal of preparing them for submission to TFSC. In alignment with the mission of TFSC and the ATLC community’s core interests, the workshop will be organized as the following 90-minute programme.

• Welcome and Opening Remarks (15 minutes) o Diana Hicks, School of Public Policy, Georgia Tech  A welcome address from Georgie Tech o Mei-Chih Hu, Editor-in-Chief of TFSC  A short introduction to TFSC, and an overview of the workshop goals, structure, and agenda.

• Panel Discussion: A Conversation between ATLC and TFSC (70 minutes) o Panellists  Diana Hicks - Professor, School of Public Policy, Georgia Tech  Caroline Wagner - Professor, John Glenn College of Public Affairs, the Ohio State University; TFSC Editorial Board Member  Mei-Chih Hu - Professor, Institute of Technology Management, National Tsing Hua University; TFSC Editor-in-Chief  Yi Zhang (Moderator) - Associate Professor, Australian Artificial Intelligence Institute, University of Technology Sydney; TFSC Executive Editor o Session Structure  Visions of ATLC and its interplays with TFSC, Diana Hicks (10 minutes)  TFSC scope, expectations, and how to craft manuscripts aligned with TFSC, Mei-Chih Hu (10 minutes)  From the perspective of TFSC editorial board, reviewers, and readers, Caroline Wagner (10 minutes)  Discussion: How to develop theoretical contributions, frame research questions, and position work bridging the interests of ATLC and TFSC. (20 minutes)  Q&A (20 minutes)

• Closing Remarks and Next Steps (5 minutes) o Yi Zhang, TFSC Executive Editor  Summary of the workshop, key takeaways, and plans for post-workshop follow-up

ABSTRACT. I. Introduction

Artificial intelligence (AI) has potential to transform economies, particularly low- and middle-income countries (LMICs), where technological advancements can contribute to achieving Sustainable Development Goals (SDGs). However, quantifying AI adoption capability in LMICs remains difficult due to limited data and the nascent stage of AI ecosystems in these regions. Building on the concept of absorptive capacity, this study proposes a novel framework to forecast AI adoption capability in LMICs. Absorptive capacity, initially conceptualized at the firm level, describes a firm’s ability to learn from external knowledge inflows and convert that learning into productive outcomes.(Cohen and Levinthal 2000, Strategic Learning in a Knowledge Economy). Khan extended this framework to LMICs’ contexts (Khan 2022, Structural Change and Econ Dynamics), emphasizing that LMICs can act as learning nations in a globalized world. In their latest work, authors further theorized this concept in Low Income Countries (LICs) for AI adoption (Khan, Umer, and Faruqe 2024; Humanities and Social Sciences Communications). Building on previous work, this research forecasts AI adoption capability using absorptive capacity as the theoretical lens, focusing on external inflows (e.g., high-tech imports) and domestic readiness factors (e.g., internet penetration and human capital).

Initially, we forecast AI capability in LMICs based on external knowledge inflows, using proxies such as imports of AI-related hardware (e.g., computers, and other technologies). This provides a baseline forecast of AI potential derived from external resources. In the next phase, this analysis is supplemented by considering domestic factors that influence a country’s ability to internalize and effectively utilize these inflows. Key factors include internet infrastructure, human capital (such as the education and technical skills of the workforce), and institutional readiness. By integrating both external inputs and domestic abilities, we produce a more nuanced forecast of AI capability that reflects external opportunities and internal readiness. This two-step approach allows us to model AI potential in LMICs even with limited direct data on AI development.

This study extends the absorptive capacity framework to AI adoption in LMICs, providing a novel lens to understand technological diffusion in resource-constrained settings. It offers a methodology to forecast AI adoption capabilities, aiding policymakers and development organizations in prioritizing AI investments. By aligning external inflows with domestic readiness, the study addresses global digital inequities, fosters economic development, and supports LMICs' meaningful engagement in the global AI ecosystem.

II. Research Questions This study seeks to address the following key question(s):

1. How can absorptive capacity be operationalized to model and forecast AI adoption capability in LMICs? This entails the following two sub-questions: i. What role do external inflows, such as high-tech imports, play in enabling AI adoption in LMICs? ii. How do domestic readiness factors, including internet infrastructure, STEM education, and governance quality, interact with external inflows to influence AI adoption?

III. Research Hypotheses Based on the research questions, this study tests the following two hypotheses:

i. Countries importing more AI-related inflows will exhibit higher AI capability. ii. AI-related inflows interact with domestic readiness factors to enhance AI capability.

IV. Methodology To answer the questions and test our hypotheses, we employ a robust time series forecasting approach using quantitative methods:

1. Data Sources, time period and sample: o External Knowledge Inflows: High-tech imports, specifically AI-related high-tech (ICT) goods (6-digit HS code level data), sourced from UNCTAD and BACI (CEPII) databases. o Domestic Readiness Factors: Internet penetration, STEM education enrollment, and institutional readiness indicators, Patent Data from the World Development Indicators (WDI) and other international databases such as patent data. o Time Period: 12 years (2010–2022) to capture trends and technological progress over time. o Sample: LMICs as defined by the World Bank in FY 2022-23.

2. Empirical Approach: o Time Series Forecasting:  Univariate Forecast: Analyze trends in high-tech imports as the baseline measure for AI-related hardware inflows.  Multivariate Forecast: Incorporate domestic readiness factors to model their interaction effects with high-tech imports. o Statistical Models:  Autoregressive Integrated Moving Average (ARIMA) models for baseline forecasts of high-tech imports.  Vector Autoregression (VAR) models to capture interdependencies between high-tech imports and readiness factors.  Dynamic panel data models (e.g., System GMM) to exploit temporal variation and address endogeneity o Interaction Effects: Model the interaction between high-tech imports and domestic readiness factors (e.g., high-tech imports × internet penetration) to assess how these factors collectively influence AI adoption capability.

3. Validation: o Forecasted capabilities will be compared against real-world trends in AI-related outputs, such as publications, patents, and startup activity, to ensure robustness.

V. Preliminary analysis:

V.1. AI-related Hardware:

Using our own understanding and, subsequently, leveraging generative AI, we identified 23 categories of AI-related hardware from UNCTAD and BACI (CEPII) databases.

V.II. Import data:

From the imports data (BACI (CEPII)), it is evident that most LMICs have significantly higher imports than exports of AI-related hardware. However, countries like Vietnam and the Philippines, exhibit higher AI-related hardware exports than imports, likely due to their growing roles as emerging production hubs for high-technology products. Exploring trends in these and other countries would provide more definitive insights into whether AI-related hardware imports can serve as a reliable proxy for trade-based measurement of AI adoption.

V.IV. Some findings from analyses:

Pending execution of our empirical approach, our analysis may suggest that AI-related imports serve as a robust proxy for AI advancement. We will also know how much readiness factors mediate the impact of external inflows on AI adoption capability. For example, countries with higher internet penetration and better STEM education systems may exhibit stronger utilization of imported AI-related technologies. Without adequate readiness, the potential benefits of external inflows perhaps could be diminished. Furthermore, by analyzing historical trends and interaction effects, we anticipate projecting that LMICs with higher absorptive capacity are likely to achieve significant AI adoption capability by 2030.

VI. Conclusion

This study provides a forward-looking framework to forecast AI adoption capability in LMICs, emphasizing the interplay between external knowledge inflows and domestic readiness. By operationalizing absorptive capacity at the LMIC-national level, it offers actionable insights for fostering AI-driven growth in resource-constrained settings, advancing digital equity, and promoting sustainable development.

ABSTRACT. The importance of innovation ecosystems and technology upgrading are emphasised in the catch-up literature. Yet, there is a limited evidence-based understanding of the interaction between technological upgrades and innovation ecosystems in contexts with persistent weaknesses in technology upgrading and low investment in research and development (R&D). Using pooled data from the two waves of 2007 and 2020 of the World Bank’s Enterprise Survey (WBES), this paper constructs novel multidimensional indicators of technological upgrades and innovation ecosystems. Estimating structural models, the paper examines the effect (direct and indirect) of the self-constructed technology upgrading and innovation ecosystem indices on the performance of productivity and export of South African firms. The findings show a positive complementary effect of technology upgrading and innovation ecosystems on labour productivity. On the contrary, the results suggest that technology upgrading and innovation ecosystems matter differently for export performance. The findings have implications for innovation policy and for firm-level strategies for innovation and technology upgrading towards enhanced performance.

ABSTRACT. 1. Background and rationale Does the capability to create economic performance steadily increase in a country with more excellent capabilities in creating innovative outcomes from R&D investments? As Romer (1990) points out in his endogenous growth theory, modern economic growth is achieved through continuous R&D investments, such as the development of new technologies and new ideas. In particular, the more active a country is in innovation through continuous R&D investments, the more likely its corresponding economic performance will steadily increase (Hasan and Tucci, 2010). R&D investments are an important foundation and key determinant of steady economic growth and productivity gains (Griliches, 1979; Furman et al., 2002; Romer, 1990). However, no clear consensus has yet been reached when examining the results of existing empirical analyses (Guloglu and Tekin, 2012; Gumus and Celikay, 2015; Pala, 2019; Hammar and Belarbi, 2021). Such mixed results are likely due to a lack of use of cross-country panel data that can identify the dynamics of national capabilities over a long period of at least 20 years, as well as the application of analytical models that can capture the nonlinearity between innovation and economic growth. This study aims to understand the relationship between innovation and economic growth from the perspective of the national innovation system (NIS) and Resource-based View (RBV) on this delay (Castellacci & Natera, 2013; Dutta et al., 2004). No matter how much R&D investment is made, the ability to convert such R&D investment into innovation results, that is, R&D capability, may be low due to a lack of skilled research personnel or insufficient research infrastructure. Furthermore, even if R&D capability is superior to other countries, if it is not supported by economic capability that translates into economic performance, a lag between innovation and economic growth may occur, as suggested by the Swedish paradox. This lag phenomenon can be approached from the NIS and RBV. According to these arguments, the total capability of the NIS can be classified into R&D capability and economic capability (Lu et al., 2016). R&D capability can be understood as a concept corresponding to innovation, and economic capability can be understood as a concept corresponding to economic growth. R&D capability can mature under the support of economic capability, and vice versa; it can also hold. The co-evolution process of these two capabilities can be nonlinear due to diminishing returns to R&D investment scale and the declining imitation effect of technology-following countries on technology-leading countries over time (Castellacci and Natera, 2013; Coccia, 2018). In this study, we empirically demonstrate that economic capability does not necessarily grow linearly as R&D capability increases. This perspective can contribute to broadening our understanding of the relationship between innovation and economic growth.

2. Methods To address the research question, we constructed a balanced panel dataset encompassing 52 countries over 24 years (1995–2018), including 37 OECD countries and 15 developing nations, including China, Indonesia, and Thailand. We utilized the Multiple Imputation method to handle missing data commonly encountered at the country level. The countries were classified into two groups—technology leaders (15 countries) and technology followers (37 countries)—based on per capita GDP and innovation activity, as measured by U.S. registered patents. Using this dataset, we applied a two-stage approach with the Non-oriented Return-to-Scale Dynamic Network Slack-Based Measure of Super-Efficiency (DNS-SBM) model, which integrates super-efficiency into the framework developed by Tone and Tsutsui (2014) as a non-oriented, variable returns-to-scale efficiency model. This model captures the network structure of NIS, allowing us to analyze their subcomponents rather than treating them as a single “black box.” The DNS-SBM model offers three key benefits: first, it incorporates super-efficiency, addressing monotonicity issues among decision-making units; second, it reflects carry-over effects, capturing the cumulative and delayed effects of R&D investments; and third, it can track long-term changes in the subcomponents of total NIS capability. For comparison, we also employed the Non-oriented Return-to-Scale Window Network Slack-Based Measure of Super-Efficiency (WNS-SBM), which combines super-efficiency and Window-DEA based on the model developed by Tone and Tsutsui (2009). To analyze the impact of total NIS capability on the per capita real GDP growth rate, we utilized the per capita real GDP growth rate as the outcome variable. We examined the determinants of the per capita real GDP growth rate across the total NIS capability scores from the DNS-SBM model and economic growth-related control variables as independent variables. Additionally, to identify whether catch-up effects or technological advancements drive productivity improvements by country type, we utilized the Global Malmquist Productivity Index (GMPI) developed by Pastor and Lovell (2005).

3. Results The DNS-SBM model revealed that the total NIS capability for technology leaders (0.636) was significantly higher than for technology followers (0.506), with the Mann-Whitney test confirming statistical significance at the p<0.05 level. The results highlight the importance of accounting for carry-over effects in measuring the total NIS capability. When analyzing the trends in the subcomponent capabilities that comprise the total NIS capability from 1999 to 2018, the DNS-SBM model revealed a steady increase in R&D capability, rising from 0.260 in 1999 to 0.628 in 2018. However, economic capability declined from 1.505 in 1999 to 0.914 in 2018. Notably, while R&D capability for technology-following countries increased, their economic capability fell more sharply than that of technology-leading countries. This trend suggests a diminishing imitation effect, traditionally viewed as an advantage for latecomers. The findings of the Global Malmquist Productivity Index analysis further validated this trend.

4. Significance and implications Policymakers should recognize that an increase in R&D capability does not necessarily lead to a corresponding increase in economic capability. Adopting an integrated approach that connects R&D and economic capability is essential to converting innovation outputs into economic growth. This approach requires institutional mechanisms to support technology commercialization, which can help transform innovation outputs into economic added value. This study shows that the total NIS capability and institutional factors, such as venture capital utilization, statistically significantly impact the per capita real GDP growth rate. Furthermore, policymakers should focus on managing semi-fixed input factors, such as knowledge and capital accumulation, subject to carry-over effects.

ABSTRACT. In the fields of innovation and growth, a country's technological capability has long been recognized as a key endogenous driver of its progression toward a more complex and sophisticated economy. While a substantial body of literature has examined technological capability, studies from the perspective of national innovation systems often categorize it into distinct types, such as production capability, investment capability, and innovation capability; production capacity and technological capability; know-how and know-why; and implementation capability and design capability. Many of these studies implicitly assume a sequential relationship among these types of capabilities. This research explicitly highlights this sequence, arguing that there are two critical stages of technological capability necessary for national economic development. Furthermore, it empirically investigates the timing at which transitions between these stages are required. The study utilizes trade and patent data from over 100 countries spanning the past 30 years. By applying the economic complexity framework, it links the product space and technology space and defines an innovation capability related with the product and technology space for measurement. The key findings reveal that, until a country’s economic complexity index (calculated from export data) reaches a range of approximately 0.8 to 1.0, most countries exhibit near-zero “related innovation capabilities.” However, achieving a higher level of economic complexity beyond this threshold critically depends on the development of “related innovation capabilities” derived from the product and technology space based on patent data. When this threshold is translated into GDP per capita, it corresponds to approximately USD 10,000–15,000, which is often the stage where countries like Malaysia and Mexico aim to escape the middle-income trap. The newly proposed metric, “related innovation capabilities,” provides insights into the level of innovation capability necessary for a country to transition into a more advanced economic stage. When analyzed alongside the economic complexity index, it allows for the identification of critical moments when a shift from production capabilities to innovation capabilities becomes essential. This research contributes to the literature by offering robust quantitative evidence that supports the transition tale of technological capabilities conceptualized in various ways by previous studies. Moreover, identifying the timing for such transitions underscores that a one-size-fits-all approach to science and technology policy is inadequate. Instead, it offers practical guidelines for policymakers to design tailored long-term national strategies. Although trade and patent data have inherent limitations, this approach can be extended to analyze other proxies for capabilities or applied to specific industries, such as green technology or the automotive sector.

ABSTRACT. Persistent gender disparities in academic publishing—such as citation gaps, gender-homogeneous collaboration networks, and regional imbalances in research contributions—have been well documented (Prakash et al., 2024). Female scholars, despite increasing representation in certain fields, continue to face systemic disadvantages, including lower citation rates and insufficient recognition for their contributions. These challenges persist across disciplines and regions, reinforcing structural inequalities in academia. The rise of Generative Artificial Intelligence (GenAI), particularly with tools like ChatGPT, offers a unique opportunity to disrupt these entrenched patterns by democratizing research processes. By lowering linguistic and methodological barriers, GenAI tools have the potential to expand access to scholarly production and foster inclusivity. However, their impact on existing gender disparities in academia remains unclear. This study seeks to investigate whether the adoption of GenAI serves to mitigate or amplify these inequities. In particular, our research addresses three sub-questions: (1) How has the introduction of GenAI tools affected gender-based collaboration patterns and authorship in academic publishing? (2) To what extent does GenAI influence the distribution of scholarly recognition and impact along gender lines? (3) Does the democratizing potential of GenAI tools manifest differently across disciplines and geographical regions, particularly concerning existing core-periphery disparities in academia? Through an inductive approach, this study aims to map the emerging pattern of gender inequality in academic publishing following the introduction of GenAI tools. Specifically, this research conducts a large-scale bibliometric analysis of publications from January 2021 to December 2023, sourced from the Web of Science database, encompassing STEM, social sciences, and humanities. Using computational gender inference methods (Goyanes et al., 2024) that integrate tools such as ChatGPT, Namsor, and Gender-API, the study examines the impact of GenAI on gender patterns through critical metrics, including collaboration patterns, authorship roles, recognition, and citation practices, and publishing stratification across journal tiers. The analysis also considers disciplinary and geographical variations to understand how GenAI interacts with existing hierarchies and core-periphery inequalities in global academia. This mapping exercise makes several contributions to understanding gender inequality in contemporary academia. First, it provides the first systematic documentation of how GenAI is reshaping gender disparities in academic publishing. Second, it identifies new forms of inequality beyond the simple Matthew effect that may require policy intervention. These findings provide crucial insights for policymakers and academic institutions promoting gender equity in an increasingly AI-mediated research landscape. By mapping emerging patterns of inequality, this study helps anticipate and address new forms of gender disparity before they become entrenched in academic practice.

ABSTRACT. The advancement of technology is built on the foundation of scientific progress, often referred to as "standing on the shoulders of giants." For example, the development of transistors, which are foundational to semiconductors in devices like computers and smartphones, stems from the advancements in quantum mechanics by scientists such as Max Planck and Niels Bohr. Similarly, the progress in artificial intelligence (AI) is rooted in foundational work, including Turing machines by Alan Turing and information theory by Claude Shannon. These historical cases highlight the critical role of innovation at the boundary of science and technology. This study investigates the mechanisms underlying innovation at the science-technology boundary. Using over 80 million journal articles and conference papers from the Microsoft Academic Graph and more than 7 million utility patents, we analyzed the relationship between papers and patents. We focused on patent–paper pairs (PPP) datasets (Marx and Scharfmann, 2024), which encompass shared knowledge content, author overlaps, and citation relationships between patents and papers. First, we measured the degree of disruptive innovation in papers and patents over time using the CD index (Funk and Owen-Smith, 2017). Consistent with previous studies (e.g., Park et al., 2023), we observed a decline in disruptiveness across both papers and patents over time. Notably, this decline was more pronounced for PPP papers compared to Non-PPP papers since 2000. This trend was observed only in papers, as no significant difference was found between PPP and Non-PPP patents. The sharper decline in disruptiveness of PPP papers suggests that they increasingly function as consolidating pieces of knowledge. Additional analysis ruled out confounding effects from citation inflation or team size. Second, we evaluated the impact of PPP papers on science and technology. By categorizing 16 million papers into four groups (PPP vs. Non-PPP and Consolidating vs. Disruptive), we measured the likelihood of papers in each group becoming "Hit Papers," defined as those in the top 5% of citations within their fields. PPP Consolidating papers demonstrated a strikingly higher probability (over 20%) of becoming Hit Papers, underscoring the influential role of consolidating knowledge at the science-technology interface. Further, when analyzing 1.3 million patents citing papers, PPP Consolidating papers exhibited the highest probability (14%) of being cited as Hit Papers by patents, emphasizing their dual impact on science and technology. Finally, we examined the combinatorial innovation pattern of PPP patents. By analyzing the z-scores of CPC code – Field of study pairs in the references of PPP patents, we identified a greater proportion of atypical combinations (z-scores below zero) in PPP patents compared to Non-PPP patents (Uzzi et al., 2013). This finding indicates that PPP patents are more likely to incorporate tail novelty while maintaining conventionality. Moreover, we observed that PPP patents citing highly Consolidating papers were more likely to pursue novel combinations of science and technology. This suggests that a robust scientific foundation fosters new technological innovations. In conclusion, this study highlights significant changes in innovation dynamics at the science-technology interface, with PPP papers and patents playing a pivotal role. While previous research has emphasized disruptive innovation, our findings underline the importance of consolidating innovation as a bridge between science and technology. Given the broad impact of scientific advances like AI, we propose the need for evaluation frameworks that recognize diverse forms of innovation and their far-reaching influence. These insights contribute to understanding the mechanisms of innovation and provide guidance for future research and policy development.

Reference Marx, M., & Scharfmann, E. (2024). Does Patenting Promote the Progress of Science?. Working Paper. Funk, R. J., & Owen-Smith, J. (2017). A dynamic network measure of technological change. Management science, 63(3), 791-817. Park, M., Leahey, E., & Funk, R. J. (2023). Papers and patents are becoming less disruptive over time. Nature, 613(7942), 138-144. Uzzi, B., Mukherjee, S., Stringer, M., & Jones, B. (2013). Atypical combinations and scientific impact. Science, 342(6157), 468-472.

ABSTRACT. Venture capitalists evaluate startups’ technological feasibility and commercial potential through technical documents like patents (Hsu & Ziedonis, 2013). Scientific papers often substantiate a startup’s technology under scrutiny. A Fortune article reported that following the collapse of Theranos, Luc Gervais, the founder of Theranos’ competitor, 1Drop Diagnostics, was asked by investors to provide scientific papers to demonstrate the feasibility of his company’s technology. In science-based industries, scientific papers signal a startup’s technological competence (Gittelman & Kogut, 2003).

However, false science is not uncommon in science-based industries. Its prevalence arises not from a lack of capability but from reliance on cutting-edge scientific knowledge, often marked by debate and uncertainty in the trial-and-error of scientific discovery (Pisano, 2006). When a startup’s technology relies on false science, it can severely harm its reputation and investor appeal (Osherovich, 2011).

The rapid increase in the number of retracted papers has made it imperative to evaluate the impact of false science on startups (Rosenblatt, 2016). As inventors and firms increasingly rely on science as a source of inspiration for innovation (Marx & Fuegi, 2020), understanding the consequences of retraction events on startups becomes ever more important.

In this study, we analyze the scientific citations in patents filed by startups to establish connections between the papers and the firms. Specifically, we investigate the effects on venture capital financing when the scientific papers cited in a firm’s patents are retracted. We differentiate between two types of science-patent linkages: genuinely relevant scientific foundations and superficial, window-dressing citations. Our findings emphasize the importance of startups critically evaluating the reliability of their scientific foundations, as this has a direct impact on their ability to attract investment. Notably, even window-dressing citations of false science in patents can undermine investor confidence and negatively affect a startup’s financing prospects.

Prior research found that patents serve as signals for startups to secure venture capital funding (Hsu & Ziedonis, 2013). Existing research has shown that patents citing scientific research tend to have higher technological and economic value (Krieger, Schnitzer, & Watzinger, 2023). This is because science serves as a roadmap for technological search, enabling inventors to produce higher-quality patents (Fleming & Sorenson, 2004). As a result, scientific citations are often seen as signals of patent value. Recognizing this signaling power, firms have strong incentives to include scientific publications in their patent citations to demonstrate a robust scientific foundation, even when no substantial connection exists. In worse cases, firms may be misled by false science, leading to flawed technological paths (Freilich & Kim, 2024).

When false science is exposed through retraction, it is expected that investors will reduce their support for companies relying on erroneous technology. However, an unresolved question remains: if a company’s technology is merely associated with false science, or if its patents only cite false science for window-dressing purposes, will investors still withdraw resources? In this study, we propose that the retraction of papers cited in a startup’s patents, even as window-dressing citations, can significantly and negatively affect the startup’s subsequent venture capital financing.

We further propose that the adverse effects of retraction events are magnified for firms operating with greater technological uncertainties. In high-uncertainty contexts, where technology merit is hard to assess, investors rely more on signals, and losing scientific endorsement heightens feasibility concerns, complicating funding.

We constructed the dataset through multiple sources. Venture financing data (2000–2021) for U.S. startups was collected from VentureXpert and Crunchbase. Startups were matched with patent data from the USPTO, patent-to-science citation data from the Reliance on Science (Marx & Feugi, 2020), and retraction details from RetractionWatch. Through matching at both the patent and firm levels, we ultimately identified 158 startups that had filed patents cited false science, along with 2,827 firms in the control group.

We treated the retraction of scientific papers as an exogenous shock and employed a stagger difference-in-differences method to identify the causal effects of paper retractions on startups’ financing.

We run linear regressions on startups’ annual likelihood and total amount of venture capital received. Before retraction, we find no evidence that venture capitalists differentiated between firms citing false science and those not. On average, firms citing false science received more frequent and larger VC funding than the control group. This suggests that citing false science may have been a window-dressing strategy to secure venture capital. However, once false science is exposed through retraction, it results in significant negative consequences. Our results show that, on average, the retraction of a scientific paper cited in a firm’s patent nearly halves the firm’s chances of securing venture capital and reduces its annual funding by 13.15% (about $259,000) across the entire post-retraction period. Further analysis shows that startups facing retractions during seed and early stages, when their technological uncertainty is higher, experience stronger negative impacts on fund raising activities.

Using textual similarity between patents and papers, we assessed whether citations to false science were genuine or window-dressing. We divided the sample into two groups based on median textual similarity: the high group likely represents genuine citations, while the low group corresponds to window-dressing citations. In the low group, many similarities are near zero, showing minimal overlap of substantive words between paper and patent abstracts. However, the results for the two groups show no significant difference, indicating that even window-dressing citations of false science still have a significantly negative effect.

The study extends the research on the impact of science on business sector (Gittelman & Kogut, 2003). With the ongoing surge in retracted papers, investors may increasingly face risks of funding firms whose technologies are associated with irreproducible or low-quality scientific research (Osherovich, 2011). For startups, even though adding a window-dressing science citation seems to be costless in firms’ patenting, their invalidation can impose substantial costs. Besides, the research contributes to the literature analyzing the governance role of retractions beyond scientific organizations (Freilich & Kim, 2024; Peng, Romero, & Horvát, 2022). Our results suggest that venture capitalists are sensitive to retraction events and they can use both retractions and patent scientific citations to evaluate a startup’s technological capabilities.

ABSTRACT. 1 Research Questions Universities and national research institutions(NRIs) are key players in conducting basic research. Universities primarily focus on curiosity-driven research while serving as hubs for talent cultivation. NRIs represent structured entities invested in and aligned with critical areas of national development. They embody the state’s strategic intent, engaging in organized, large-scale research activities. The two differ significantly in their strategic orientation, research organization model, and resource allocation approaches. This study aims to analyze the heterogeneity between universities and NRIs in the process of conducting basic research. The study selects the C9 League universities (C9 League), Chinese first alliance of top-tier universities, as a representative sample of universities, and research institutes under the Chinese Academy of Sciences (CAS institutes), the country's largest and most advanced national research institution in the natural sciences, as a representative sample of national research institutions. By examining General Program funded by the National Natural Science Foundation of China (NSFC) from 2009 to 2018 (a total of 30,457 projects) and their associated research publications, this study systematically analyzes the differences in research content and thematic characteristics between these two types of institutions. This study concludes that universities and NRIs follow distinct trajectories in basic research, driven by their heterogeneity, mechanisms, and influencing factors. The findings offer empirical support for optimizing resource allocation and research policies. 2 Methodologies 2.1 Data Sources The input data for this study comes from the funding data provided by Cingta NSFC for the years 2009-2018, while the output data is sourced from academic output data under Dimensions NSFC. Considering the time lag between funding input and output, we utilized data spanning 2009-2024. These data were used to analyze the activity levels of specific research topics and the structure of institutional collaboration networks. 2.2 Data Processing This study employs a Data-driven research approach to extract useful and valid information from extensive datasets. Irrelevant or low-quality data are filtered out, and external databases such as SciVal are matched and linked to construct a comprehensive General Program input-output dataset. Additionally, external indicators of Novelty and Disruptive are introduced to enrich the analytical dimensions and provide deeper insights into the data. 2.3 Indicator Construction This study establishes metrics across three dimensions: Interdisciplinary Collaboration, Topic Features, and Degree of Innovation, to comprehensively measure and analyze the research output and its characteristics. Specifically, the analysis is divided into the following seven indicators: Interdisciplinary proportion, Disciplinary count, Topic number count, Topic number diversity, Topic prominence, Percentile, Novelty and Disruptive. 3 Research Findings 3.1 Input Analysis Earth Sciences: The number of General Program approvals for CAS institutes is approximately three times that of the C9 League. This is attributed to the CAS institutes' access to more advanced experimental equipment, observation networks, and computational facilities, which support large-scale and long-term research projects. These resources also facilitate extensive fieldwork and long-term monitoring of observation station data. Life Sciences: CAS institutes have consistently received more project approvals than universities in this field. This stems from the need for larger, specialized research teams to conduct resource-intensive, high-cost experiments, which heavily depend on deep expertise and resource advantages. Mathematical and Physical Sciences & Information Sciences: The number of the C9 League General Program approved far exceeds that of CAS institutes, which is related to the university's management system. This can be linked to the universities' organizational structure. As single legal entities, universities can effectively coordinate research across faculties and respond flexibly to market demands, enhancing their ability to translate scientific achievements into applications. In contrast, although the Chinese Academy of Sciences is large in size and rich in scientific research resources, its various research units are independent legal entities, which makes it difficult to coordinate and integrate innovative resources. 3.2 Output Analysis In most academic disciplines, the characteristics of basic research projects differ significantly between the C9 League and CAS institutes. Research conducted by the C9 League is characterized by a larger number of topics, greater dispersion, and higher levels of interdisciplinarity and disruptiveness. These traits reflect the universities’ emphasis on free exploration and the pursuit of academic frontiers. In contrast, research at CAS institutes is more focused, with topic clusters that tend to be more niche. This aligns with CAS institutes' mission of conducting mission-oriented research in support of national strategies. 3.3 Results and Conclusions 3.3.1 Results The basic research directions of universities and NRIs demonstrate a high degree of alignment with National science and technology strategy. Their differences in disciplinary focus highlight a strong role complementarity. CAS institutes, leveraging its advantages in equipment, funding, and team resources, has achieved rapid development in fields such as Earth Sciences and Life sciences. In contrast, the C9 League, with their flexible management systems, has shown greater adaptability in Mathematical and Physical Sciences and Information Sciences. This complementarity has significantly supported the comprehensive development of Chinese basic research and provides a valuable reference for optimizing the functional division and collaborative mechanisms between universities and research institutions across different disciplines. Universities prioritize curiosity-driven research, emphasizing interdisciplinary integration and publication impact, while NRIs focus on mission-oriented, application-driven research addressing national needs with field-specific specialization. Together, they form a complementary dual-track system, enriching China's basic research landscape. 3.3.2 Conclusions In terms of policy design, we should strengthen support for curiosity-driven research in universities, and provide a more flexible policy environment for NRIs to carry out task-oriented research; in terms of resource allocation, we should consider the differentiated advantages of universities and NRIs, optimize the funding allocation mechanism, and promote the balanced and coordinated development of basic research. This study highlights key differences in the research content and characteristics of universities and NRIs in basic research. Future studies could expand the data sources, subjects, and scope to further explore their connections and differences, offering insights for the long-term development of basic research and the innovation ecosystem.

ABSTRACT. Due to the highly advanced aerial transportation technology, commercial operations recently began to turn up in several areas including New York, Mexico City, and Dubai. As unmanned aerial mobility (UAM) offers fast and quiet travel, it is expected to expand in use in both urban and rural areas. However, whether the public is favorable to this technology remains questionable. As the consideration of the public acceptance of emerging technologies before their diffusion is necessary, this study attempts to conduct a multivariate analysis on UAM acceptance in South Korea. South Korea is selected as a case study due to the recent attempts of the Korean government to introduce UAM for cities that suffer from high traffic (Future Drone Traffic Manager, 2020; UAM Team Korea, 2021). Preliminary research that focused on the public acceptance of the technology used the Technology Acceptance Model (TAM) measuring user’s perceived usefulness and perceived ease of use, and adding the trust variable as people are unconfident in emerging technologies (Chancey, 2020; Johnson et al., 2022; Kim et al., 2023). In addition, they usually focused on unmanned and automated vehicles that were devised to scale up operations sacrificing safety (Chancey, 2020; Johnson et al., 2022). However, citizens remain reluctant to ride a vehicle without a pilot, and the Korean government is planning to require a pilot to board the UAM for its operation (UAM Team Korea, 2021). Therefore, studies that focused on unmanned UAMs provide insufficient information for immediate policy implications. Moreover, preliminary studies tend to focus only on the perspective of potential users, which excludes the non-users out of the scope who might feel discomfort with the commercialization of UAM. Therefore, this study focuses on the acceptance of piloted UAM considering both users and non-users. This study uses the Extended Unified Theory of Acceptance and Use of Technology (UTAUT2) model, which is an extended model of UTAUT that combines eight technology acceptance models including Theory of Reasoned Action (TRA), TAM, and Theory of Planned Behavior (TPB) (Venkatesh et al., 2012; Alomary & Woollard, 2015). To study the public acceptance of UAM, an online survey with 1800 South Korean respondents was conducted. The respondents were grouped into two depending on their intention to use UAM, namely, the users and non-users. They were asked about their sociodemographic information that is related to the acceptance of technology such as age (Arning & Ziefle, 2007), and gender (Gefen & Straub, 1997; Venkatesh & Morris, 2000) in addition to behavioral intention (BI), performance expectancy (PE), effort expectancy (EE), social influences (SI), facilitating conditions (FC), hedonic motivation (HM), price value (PV), and habit (Venkatesh et al., 2012). Additionally, they were asked whether they would feel safe when using UAM (SWU) and when they were near (SWN). The survey respondents were given a short introduction to read about air taxis and their performance before answering the questions. Structural equation modeling is used to analyze social acceptance towards UAM in South Korea. The results indicate that performance expectancy, social influence, hedonic motivation, price value, and safety when near are key predictors of behavioral intention to use UAM. Notably, these variables were significant in evaluating the intention of potential users. However, only hedonic motivation and price value remained as important predictors in the non-user group. This study contributes to the literature on social acceptance towards UAM by introducing an additional dimension, dividing the respondents into two groups. Current studies tend to focus solely on future users, which provides limited scope and understanding, as the use of UAM may affect non-users’ quality of life, residing location, safety, and privacy. Understanding the key determinants of the acceptance of both groups serves as a foundational step for the discussion between the public and the UAM industry. Furthermore, this study may provide a reference for technology developments to enhance public acceptance on UAM, addressing key determinants of users and non-users.

Reference Alomary, A., & Woollard, J. (2015). HOW IS TECHNOLOGY ACCEPTED BY USERS? A REVIEW OF TECHNOLOGY ACCEPTANCE MODELS AND THEORIES. Arning, K., & Ziefle, M. (2007). Understanding age differences in PDA acceptance and performance. Computers in Human Behavior, 23(6), 2904–2927. https://doi.org/10.1016/j.chb.2006.06.005 Chancey, E. T. (2020). Effects of Concepts of Operation Factors on Public Acceptance and Intention to Use Urban Air Mobility (UAM) – Trust and Technology Acceptance Modeling. Air Mobility. Future Drone Traffic Manager. (2020). 도시의 하늘을 여는 한국형 도심항공교통(K-UAM) 로드맵 [Urban Air Mobility(K-UAM) Roadmap opening the city sky]. https://www.molit.go.kr/USR/NEWS/m_71/dtl.jsp?id=95083976 Gefen, D., & Straub, D. (1997). Gender Differences in the Perception and Use of E-Mail: An Extension to the Technology Acceptance Model. MIS Quarterly, 21, 389–400. https://doi.org/10.2307/249720 Johnson, R. A., Miller, E. E., & Conrad, S. (2022). Technology Adoption and Acceptance of Urban Air Mobility Systems: Identifying Public Perceptions and Integration Factors. The International Journal of Aerospace Psychology, 32(4), 240–253. https://doi.org/10.1080/24721840.2022.2100394 Kim, Y. W., Lim, C., & Ji, Y. G. (2023). Exploring the User Acceptance of Urban Air Mobility: Extending the Technology Acceptance Model with Trust and Service Quality Factors. International Journal of Human–Computer Interaction, 39(14), 2893–2904. https://doi.org/10.1080/10447318.2022.2087662 UAM Team Korea. (2021, September 28). 한국형 도심항공교통(K-UAM) 운용개념서 1.0 [Korean Urban Air Mobility (K-UAM) Management Concept Description 1.0]. https://www.molit.go.kr/USR/NEWS/m_71/dtl.jsp?lcmspage=1&id=95086041 Venkatesh, V., & Morris, M. G. (2000). Why Don’t Men Ever Stop to Ask for Directions? Gender, Social Influence, and Their Role in Technology Acceptance and Usage Behavior. MIS Quarterly, 24(1), 115–139. https://doi.org/10.2307/3250981 Venkatesh, V., Thong, J. Y. L., & Xu, X. (2012). Consumer Acceptance and Use of Information Technology: Extending the Unified Theory of Acceptance and Use of Technology (SSRN Scholarly Paper 2002388). Social Science Research Network. https://papers.ssrn.com/abstract=2002388

ABSTRACT. With recent AI advancements, public awareness of AI-related harms—like bias, discrimination, and privacy violations—has increased, prompting efforts to align AI systems with public values. Various bodies have developed frameworks that require AI systems to conform with values such as transparency, explainability, and fairness; while major AI developers (‘Big Tech’) have introduced their own responsible AI principles. However, these values and principles are often critiqued as vague and too high-level (Schiff, 2023) raising questions about how they should be operationalized.

This challenge is not new in the field of responsible research and innovation (RRI). For instance, Tabares et al. (2022) documented the diffused implementation of RRI across eight programs of the European Commission’s Horizon Framework Programme for Research and Innovation (H2020), highlighting challenges in moving from theory to practice. Christensen et al. (2020) also explored how the perceptions and operationalizations of RRI varied among five different organization types, namely public funding agencies, private research foundations, universities, private companies, and civil society organizations, across different countries. Such studies reveal that values and principles as those promoted by Big Tech and other authoritative bodies do not smoothly translate from into practice, undergoing significant modification by different stakeholders. However, systematic understanding of how these changes occur is still limited.

To contribute to this understanding, we examine the case of the Georgia Artificial Intelligence in Manufacturing (Georgia AIM) Project, a $65 million federally funded initiative to integrate AI into Georgia’s manufacturing economy. Our study focuses on equity as a project value, asking: 1. How is equity being operationalized in the Georgia AIM project? 2. What factors influence equity’s operationalization in this project?

To answer, we collected data from both primary and secondary sources. For the primary sources, we facilitated a workshop on equity for manufacturing extension agents involved in the project (see workshop report: https://doi.org/10.31235/osf.io/ujfsq) and conducted 24 semi-structured interviews with project leaders. For the secondary sources, we reviewed project documents including the Economic Development Administration [EDA]’s call for proposals [CFP], and the grant proposals, namely the overarching project narrative for Georgia AIM and the project narrative for each of the 8 subprojects under Georgia AIM. We employed thematic analysis to analyze the data using NVivo qualitative data analysis software.

Focusing on 3 aspects of equity, namely the what, whom, and how of equity, the thematic analysis uncovered patterns that answer our first research question. We discovered that the object (what) of equity expands during operationalization. Equity moves from being a narrow economic conceptualization influenced by the sponsor (EDA)’s definition, to including socio-economic considerations in response to environment-specific needs of the projects and other broader aims. In contrast, the subject (whom) of equity narrows as it moves from conceptualization to operationalization, with the identification of beneficiaries becoming more specific during implementation. Lastly, we found that the how of equity becomes more varied and intensified during operationalization along the four pathways of processes, programs, places, and projects. To answer the second research question, the thematic analysis further identified 8 factors as the determinants of how equity is being operationalized in the Georgia AIM project. These are logistical, contextual, political, institutional, metrological, technological, resource-based, and personal factors. The aggregate impact of all the factors combined leads to the variations observed in the implementation of equity and determines how equity is operationalized. This, however, is not an identification of a strict causal relationship between each factor and the variation observed in the operationalization of equity. Rather, it is a systematic categorization of the influences on the operationalization of equity, to serve as a heuristic for discovering specific causes of variations in implementing equity. The factors discovered in this study are not comprehensive or exhaustive and could be expanded by further research.

Our findings have two main implications. First, we provide empirical evidence on how equity, as a pillar of public administration, evolves in large-scale public innovation and economic development projects. We show how equity changes from conceptualization through implementation and offer a pragmatic heuristic to aid project leaders in successfully implementing equity. While this is a single case study with limited generalizability, our study provides an initial understanding of how equity changes helpful for project leaders to anticipate the identified patterns of change and devise plans for addressing important equity determinants that may be crucial for project success.

Second, our study advances the RRI literature. While many RRI projects aim to incorporate equity, few studies have examined its real-world dynamics. In providing evidence that principles like equity are seldom static during implementation even within the same project, our study provides useful insights that are especially relevant to one of the areas of responsible innovation that has received much attention lately, responsible AI. With the proliferation of AI principles and ethical frameworks by organizations and different stakeholders, this study provides evidence that these principles will not translate directly as they have been conceptualized but will instead be subject to many variations and evolutions during their operationalization. By understanding how principles like equity translate and the factors that should be considered during their implementation, organizations and other stakeholders may be able to better realize the goals of responsible AI.

References Christensen, M. V., Nieminen, M., Altenhofer, M., Tangcoigne, E., Mejlgaard, N., Griessler, E., & Filacek, A. (2020). What’s in a name? Perceptions and promotion of responsible research and innovation practices across Europe. Science and Public Policy, 47(3), 360-370. https://doi.org/10.1093/scipol/scaa040 Schiff, D.S. (2023). Looking through a policy window with tinted glasses: Setting the agenda for U.S. AI policy. Review of Policy Research, 40(5), 729-756. https://doi.org/10.1111/ropr.12535 Tabarés, R., Loeber, A., Nieminen, M., Bernstein, M. J., Griessler, E., Blok, V., Cohen, J., Honigmayer, H., Wunderle, U., & Frankus, E. (2022). Challenges in the implementation of responsible research and innovation across Horizon 2020. Journal of Responsible Innovation, 9(3), 291–314. https://doi.org/10.1080/23299460.2022.2101211

ABSTRACT. Short Abstract We combine scientometrics and generative AI (ChatGPT) to analyze the intersection of AI and green technologies. Leveraging technological trajectories built using the entire EPO data, the study identifies current common themes and projects future directions of twin (AI and green) technologies. Generative AI summarizes patent information, revealing technological paradigms and their expected future evolution. Initial findings highlight key areas like automation, environmental management, and IoT integration. In ongoing work we are expanding the analysis, enhancing replicability, and comparing results with other NLP methods for a more robust representation of past and future technological developments.

Extended abstract The intersection of artificial intelligence (AI) and green technologies represents a crucial frontier in addressing the climate emergency. This paper studies the dynamic landscape of AI-related green innovations, aiming to map current trends and project future trajectories of technological development. Our methodology combines established scientometric analysis with the emerging capabilities of generative AI, leveraging a comprehensive database of technological trajectories constructed from the all inventions in the European Patent Office (EPO) data.

We focus on the convergence of AI and green technologies within our trajectory database that covers all inventions in the EPO. Technological trajectories, represented as chains of patents linked by citation relationships, offer valuable insights into the evolutionary pathways of these inventions. We employ generative AI, specifically ChatGPT, to synthesize information extracted from patent titles and abstracts, generating concise summaries for each technological trajectory.

This AI-driven summarization approach allows us to efficiently analyze a substantial number of trajectories. Our initial dataset comprises approximately 43,000 trajectories identified as relating to both green technologies (classified using Y02 and Y04S CPC codes) (Nomaler and Verspagen, 2019) and digital technologies (as identified in our earlier work (Menendez De Medina et al, 2023; Prytkova et al 2024)). We then refined this set, focusing on 950 trajectories that explicitly mention "artificial intelligence" or "machine learning" within their titles or abstracts. From this refined set, a random sample of 32 non-overlapping trajectories was selected for in-depth analysis, ensuring that no single patent was represented in more than one trajectory.

For each trajectory within this sample, we provided a structured prompt to ChatGPT, requesting: (1) a descriptive summary of the trajectory’s technological development, (2) a reasoned prediction of potential future directions for the technology, and (3) identification of closely related technologies that could contribute to these future advancements. We provided to ChatGPT a data record containing title, abstract, priority year, and the position in the trajectory. ChatGPT's responses were formatted to clearly distinguish between these three requested components.

Building upon these individual trajectory summaries, we further utilized ChatGPT to identify common technological "paradigms" emerging across the set of trajectories. These paradigms represent overarching themes or fundamental technological building blocks that characterize groups of related trajectories. We conducted separate analyses for past and future paradigms, providing ChatGPT with the descriptive summaries of the trajectories (1) for the former and the predicted future directions (2) for the latter. This process resulted in the identification of eight distinct past paradigms, including categories such as "Automation and Control Systems," "Environmental Management and Sustainability," and "Energy Productino and Management." For future paradigms, ChatGPT identified six key paradigms, such as "Integration with IoT and Smart Systems," "Enhanced AI and Machine Learning Integration," and "Sustainability and Energy Efficiency."

To visualize the relationships between these paradigms and the underlying trajectories, we constructed a tripartite network. In this network, the 32 trajectories serve as the connecting links between the past and future paradigms, as each paradigm is characterized by the set of trajectories it encompasses. We employed VOSviewer (Van Eck and Waltman, 2010) to map this tripartite network, providing a graphical representation of the patent landscape for green technology and AI derived from our ChatGPT-assisted analysis. The resulting network map revealed a central cluster of broadly applicable future paradigms, surrounded by more specialized clusters of trajectories and past paradigms. These specialized clusters corresponded to areas like mobility (with a focus on emission control), wind turbine technology, industrial automation, and environmental monitoring, offering a structured overview of the key areas of innovation within the AI-green technology domain.

This pilot study provides evidence for the potential of combining traditional scientometric methods with the capabilities of generative AI to gain insights into complex technological landscapes that may be difficult to describe with standard labelling methods and topic modelling.

The current study is based on a relatively small sample of trajectories. Moreover, the reliance on ChatGPT introduces inherent challenges to replicability. We are currenlty addressing these limitations by scaling up the analysis to encompass a significantly larger number of trajectories, ultimately aiming to cover the entire dataset of 43,000 trajectories related to green and digital technologies. Furthermore, we are implementing this work using generative AI systems that can be replicable, such as Llama, offering greater control over parameters. Finally, we will compare and contrast our generative AI-driven findings with results obtained through more established NLP techniques, such as sentence embeddings and semantic similarity analysis, to validate the insights generated by AI and ensure the reliability of our conclusions.

ABSTRACT. Emerging digital automation technologies such as AI can have several impacts on the future of work, depending on the extent to which they can replace, complement, or create new tasks. For example, the automation and computerization waves of the 1990s and early 2000s have been impacting occupations with high content of routine tasks, especially manual (Goos and Manning, 2007; Acemoglu and Autor, 2011; Goos et al., 2020). AI and related technologies, though, are being developed to perform tasks that were commonly referred to non-routine (Frey and Osborne, 2017; Nedelkoska and Quintini, 2018) and abstract tasks (vom Lehn, 2020), which use human observation and judgement that is not standardised (Savona et al., 2022).

In this paper, we combine views from a thousand experts from science, technology, business, civil society, and policy to make predictions about which tasks different emerging digital automation technologies will perform by 2030.

To do so, we first implemented an articulated real-time Delphi survey. The survey asked experts from academia, industry and civil society to identify the three technologies most likely to be used by 2030 from around 600 emerging automation technologies, grouped into eight families — AI & Intelligent Information Systems, Robotics, Computing, Networking, Data Acquisition, Data Management, User Interface and Additive Manufacturing Technologies — and several subfamilies (see Olivera et al, 2004).

Next, we asked what tasks these technologies will be able to perform by 2030. We first asked to select among all the O*NET Broad Work Activities — such as Getting Information, Making Decisions and Solving Problems or Handling and Moving Objects. We then use O*NET Detailed Work Activities to obtain more fine-grained knowledge about the technology’s capability to perform specific tasks. Once they had connected technologies with detailed work activities, survey participants were asked to assess to which extent technology-work activity pairs complement or substitute one another and to which extent the technology codifies or enhances the efficiency of such respective work activity.

Preliminary results show that the most relevant emerging automation technologies (largely driven by AI) will mostly perform Information and Data Processing activities, and activities related to Reasoning and Decision Making. At the other end, the Broad Work Activities least selected relate to Communicating and Interacting, and Coordinating, Developing, Managing and Advising.

Automation technologies, though, vary in the kind of work activities they can perform based on their technical features, applications, and industries of adoption. To assess the exposure of work activities to each technology family, we compute the ratio between the relative frequency of each BWA by technology family and the average frequency for all technology families. The results confirm the general purpose of AI emerging technologies. With the exception of manual activities such as handling and moving objects, and performing general physical activities, AI technologies can perform most other activities. The only other activities in which AI is less relevant to human are those related to monitoring processes and resources, and operating vehicles: these are tasks performed mainly by IoT and related networking technologies.

To understand whether the top emergent technologies will more likely replace, complement, codify tasks or make them more efficient, we asked survey participants to assess each of the following statements on a 5 points Likert scale: 1. By 2030, this technology will replace humans in carrying out the following task; 2. By 2030, the following task will be sufficiently standardized (or routinized) that they can be carried out automatically by this technology; 3. By 2030, the knowledge/information to undertake the following task will be sufficiently codified and written in protocols for this technology to carry it out; 4. By 2030, this technology will make this task more efficient.

For all eight technology families, participants strongly agree with the high potential that emerging technologies have in increasing the efficiency of performing the task. Responses differ substantially when assessing if the technology will complement or substitute humans. On the one hand, most experts agree on the complementarity between digital automation technologies and humans in carrying out a given work activity across all families. On the other hand, experts’ opinions diverge when assessing machine-human substitution. The level of consensus varies across technology families. Experts in Robotics and AI think that these technologies are more likely to substitute humans in the corresponding tasks than, for instance, Computing, Additive Manufacturing, and User Interface technologies. In Robotics, this is mainly the case for Manipulative robots, Software/virtual robots, and Smart Materials. In the case of AI, it is Robotics applications and Computer Vision, where the largest share of experts consider that the technology may substitute humans.

Building on our unique survey results, we revise and complement existing measures of occupational exposure (Arntz et al., 2017; Nedelkoska and Quintini, 2018). Our exposure measure considers the task content of each occupation, the relevance of each task to a given occupation, and the individual exposure of each task to automation. Moreover, our measure is forward-looking: it is based on exposure to technologies that are more likely to be of high relevance in the near future, increasing the measure’s reliability. Finally, by considering different forms of technology-labor interaction (i.e., whether a given technology is complementary or substitutable to a given task), we address nuances that bring valuable information for decision-making: workers in occupations composed of many tasks highly exposed to automation substitution can be more affected than workers in occupations where tasks are highly complementary to technologies.

Our finding suggest that we may expect a different reshuffle of jobs with respect to the polarisation of labour observed with ICT. The most exposed to emerging technologies are clerical support workers, technicians and professionals (as also shown by complementary work in Prytkova et al. (2024)). These include routinized mid skilled occupations such as secretaries, but also highly educated ICT technicians and higher education teachers. Low skilled elementary and craft and related trade workers are the least exposed. However, differently from previous work, we distinguish between substituting and complementing tasks, and show that high skilled professionals will be using AI and information technologies to complement their work.

ABSTRACT. Background and Rationale

Policymakers in many U.S. states are making significant investments in the closely related goals of (1) increasing the rates at which individuals from low-income backgrounds are able to achieve a stable income and a family sustaining wage, (2) growing college enrollment and graduation among adult-age individuals, and (3) strengthening state economic resilience and vitality by increasing growing the share of the workforce who hold educational credentials with value to the labor market. In alignment and support of these goals, federal policymakers are making substantial investments in improving access to and completion of career and technical education (CTE) programs—particularly at the subbaccaluareate level—that are aligned with current and emerging STEM workforce needs. Unfortunately, efficient resource utilization to maximize the achievement of these goals is hampered by sizeable gaps in our understanding of the landscape of subbaccaluareate education, pathways through postsecondary education, students’ use of those pathways, and the relationships between subbaccaluareate education and workforce opportunities. One blind spot that is especially impactful due to its size and the potential for disparate outcomes for historically marginalized students is the role of for-credit stackable credential pathways.

The U.S. Department of Labor defines these stackable credential programs as “a sequence of credentials that can be accumulated over time to build up an individual’s qualifications and help them to move along a career pathway or career ladder to different and potentially higher-paying jobs.” (US DOL, 2020, p. I-3). Stackable credential programs are designed to operate as follows: A student enters community college and first earns a postsecondary certificate, which requires under two years of full-time equivalent coursework. The certificate-earner then exits the community college and works full-time in a position related to their certificate. Some time later, the certificate-earner reenrolls in college, pursuing another closely related credential that builds upon credits they have already earned and knowledge they have already mastered. Yet, emerging research demonstrates that community college students may stack credentials in myriad ways beyond policymakers’ and administrators’ original designs (Daugherty et al., 2022).

It remains unclear how students are leveraging stackable credentials to improve labor market opportunities in STEM. Said another way, we do not fully understand where credential stacking fits in students’ labor market trajectories. For instance, we do not know to what extent students are compelled to re-enroll in college and earn a second credential in a stackable sequence by an economic shock, by low earnings following their first credential, by opportunities for advancement with substantial earnings gains, or by other precipitating factors. Therefore, we ask, among students who have completed a first CTE credential, how are patterns of labor force participation, industry changes, and re-enrollment related to earning a stacked credential, , and horizontal versus vertical stacking? How do patterns of intra-field versus inter-field stacking inform our understanding of students’ use of credential stacking for labor market goals?

Methods and Data

We use administrative datasets from the Ohio Longitudinal Data Archive (OLDA), matched with earnings records from the state’s Unemployment Insurance (UI) systems and enrollment and credential records from the National Student Clearinghouse (NSC). OLDA data included all public postsecondary institutions in Ohio, including records of students’ credentials, enrollments, transcripts, and demographics.

We constructed a sample of all students who earned a short or long-term certificate from an Ohio community college between July 1, 2006 and December 31, 2013. Restricting the sample to individuals who earned their first certificate between 2006 and 2013 ensured that we could observe a full six-year time window of subsequent employment and educational behavior, prior to the onset of the COVID-19 Pandemic. Our sample includes all individuals aged 20 to 60 at the time at which they earned their first certificate in a CTE academic field5. Certificate programs in non-CTE fields remain uncommon and it is unclear the extent to which they attract working adults. The sample was further restricted to exclude those who were ever dual-enrolled in high school and an OCC, who earned multiple “stacked” credentials simultaneously with their first credential, and those students without a valid workforce ID only include students who earned their first certificate in a CTE field. These sample restrictions resulted in a final sample of 23,548 unique students.

The research was supported by a Lumina Foundation grant to the University of Michigan. We thank our partners at the Ohio Department of Higher Education and the Ohio Education Research Center for data access.

Results

These findings suggest that successful stackers tend to remain in the education system continuously or nearly continuously between their first and second stackable certificate. Most students who successfully complete stackable credential pathways exited college and entered the workforce between their first and second certificates. The vast majority were enrolled in college in the semester immediately following the award of their first certificate. This means that stacking-eligible individuals who fully exit college for any length of time rarely return to complete a second certificate. We also find that student behavior within stackable credential pathways is highly influenced by the field of study of a student’s first certificate. We do not find that changing professional fields or poor local economic conditions encourage credential stacking.

Significance

These results raise important questions about the validity of program models which propose periods of temporary non-enrollment in higher education as part of the stackable credential pathway. Such exits may, in fact, discourage degree completion and hamper student performance in stackable credential pathways. This research helps us improve our understanding of students’ actual behavior in stackable credential programs and suggests opportunities to improve our theoretical understanding and evaluation of such programs. It also helps us better understand the role and potential of stackable credentials as a tool for workforce development and retraining.

ABSTRACT. Artificial Intelligence (AI) technologies profoundly transform our social and economic systems, spreading to most fields (Agrawal et al., 2019). The growing importance of AI has made the knowledge related to those technologies a highly valuable asset for individuals when competing in the labor market. Extant studies have focused their attention mainly on the demand side of the labor market and documented the companies’ increasing demand for workers with AI training (Squicciarini and Nachtigall, 2021; Acemoglu et al., 2022; Acemoglu and Restrepo, 2019). Little is known about the supply side, i.e., those individuals with AI knowledge facing the labor market. We aim to fill this gap by assessing the impact of pursuing AI training on individuals’ career paths. As suggested by Lane and colleagues (2024), “any attempt to describe the economy-wide impact of public investment in AI would involve identifying the people at the heart of these investments’’ (page 303). In the last decade, governments, seeing educated workers as the engine of economic growth, have invested increasing funds in their higher education systems. A substantial part of those funds has been directed to train PhD students (Cyranoski et al., 2011), individuals trained at the frontier of science and technology. In this paper, we ask: Does training in AI affect PhD students’ careers? To add further insight into the debate on the privatization of AI research (Jurowetzki et al., 2021), among those who pursue a research career: Do PhD graduates with AI knowledge end up in the private sector? and What is the impact of having AI knowledge on their scientific productivity, visibility, scientific network, and international mobility? Our study considers field heterogeneity by separating fields in which AI knowledge can be easily acquired after graduation (Nelson and Winter, 1982), i.e., Computer Science, from fields in which learning AI during the PhD is expected to be a unique occasion for acquiring AI knowledge. Indeed, the theoretical, methodological, and empirical knowledge base developed during a PhD in Computer Science is similar to the one required to easily learn AI knowledge after graduation. On the contrary, for PhD graduates in fields other than computer science, e.g., a PhD in Biology, the theoretical, methodological, and empirical knowledge base developed during the PhD is far from the one required to easily learn AI knowledge after graduation. Methods To assess if AI training affects PhD students’ careers, we draw on a unique dataset covering the entire population of 46,657 French PhD graduates in STEM disciplines between 2010-2018. We leverage the national centralized archive of French PhD dissertations maintained by the Bibliographic Agency for Higher Education to extract the biographic information on PhD graduates (full name, graduation institute, defense year, supervisor name, and discipline of graduation) and the thesis text (title and abstract). We rely on the OpenAlex database to gather the bibliometric records of PhD graduates and supervisors (the published articles with their title, abstract, publication year, and author list). To reconstruct the fundraising activity of supervisors, we collect the individual grants awarded by the National Agency of Research, the main funding source for French researchers. Lastly, we use data from the European Patent Office to retrieve detailed information on patents whose inventors’ names match those of our PhD graduates, including the application filing year, the inventor’s name and country, and the applicant institution type (private versus public). We leverage access to the PhD theses to identify PhD graduates with AI knowledge as those who completed a thesis listing at least one AI keyword in the title or abstract. To identify AI keywords, we rely on both lists proposed by Cockburn et al. (2018) and the OECD (Baruffaldi et al., 2020; Calvino et al., 2023). We reconstruct careers and scientific outcomes using patent and publication data. PhD graduates who pursue a research career are those who publish at least one paper or file a patent in the 5 years after graduation. Among those graduates who pursue a research career, we identify PhD graduates working in the private sector if they are affiliated with an R&D department of a private company. To assess graduates’ scientific outcomes, we count the number of publications and patents, number of coauthors, and citations received. To assess international mobility, we use the affiliations listed in the publications to identify those who move abroad. Establishing a causal link between training in AI and career outcomes is challenging because students who developed AI knowledge during their PhD might differ from other PhD graduates. The differences between the two groups may also affect their careers after graduation. In our study, we overcome this concern and provide causal evidence by implementing a Propensity Score Matching (PSM) approach. For each PhD graduate who developed a thesis in AI, we find a ‘control’ graduate with similar characteristics but not having developed an AI thesis. We match AI graduates with similar controls considering the supervisor’s characteristics, such as mentorship experience, publication records, previous AI knowledge, connections with the private sector, and fundraising experience. We also consider student’s characteristics such as the university and discipline of graduation and the defense year. Main findings Our PSM exercise shows that AI knowledge does not significantly affect the probability of pursuing a research career for graduates, both in Computer Science and other fields. However, AI students in Computer Science are less likely to end up in the private sector than their non-AI counterparts. Having AI knowledge is detrimental to patenting activity in Computer Science, while fosters patenting activity in fields other than Computer Science. Moreover, for those students who had an AI training during the PhD, we observe for all fields a significant path dependence in continuing publishing on AI topics after graduation. Finally, we do not observe significant differences between AI and non-AI trained students when looking at their scientific productivity, visibility, and professional network. However, we observe that AI training stimulates mobility to other countries after graduation in fields other than Computer Science.

ABSTRACT. Background and rationale Since World War II, the relationship between U.S. science and policy has evolved in the face of continual societal and technological change. The social contract for science and the prioritization of basic research funding forged in the aftermath of the Manhattan Project lasted for only a brief period after the war. While scholars debate the timing and precise characterization of subsequent science policy paradigms, they describe a normative trend toward production of knowledge for societal utilization—including governance—by a wide range of actors and sites outside of academia. Facilitating these forms of knowledge creation and exchange requires new practices and entities, for which scientists and engineers typically don’t receive any training or incentivization. The establishment of new U.S. programs to engage academic researchers in policy has grown at an exponential rate with many more opportunities available in the last few years for these groups to participate in training and direct placements within government, or government affairs, offices. This recent trend suggests that an evolution of the science-policy interface may be occurring, but there is little extant data on the structure, aims, and impacts of these programs. A new database we created describes the current landscape of U.S. programs seeking to train academic researchers to engage in policy. Further, we focus on a case study of a state with a substantial number and diversity of programs to characterize the nature of these interventions: how they conceptualize audiences, activities, and impacts; and which researcher roles in policy and forms of decision-maker evidence use they address. Through these data we illustrate the scope and dimensions of these initiatives, with implications for national U.S. research and governance.

Methods To develop the database of U.S. science policy programs, we first identified data repositories such as the National Conference of State Legislatures and the National Science Policy Network, along with inputs from professional networks. We reviewed program websites for eligibility and collected additional details, subsequently contacting program leaders to verify the data. The database includes programs seeking to promote researchers’ engagement in policy processes, categorized based on activities and governance focus. We utilized the national database to identify existing science policy programs in Virginia and conducted pre-surveys and interviews with program leaders.

Results In our landscape mapping of U.S. science policy programs, we identify a substantial number (n=174) and broad diversity of programs that have emerged to engage researchers in policy processes, with the majority operating at the state level (n=99, 56.9%). Approximately half of programs at the state level are student organizations (50.5%), rounded out by government placements and fellowships (21.2%), academic certificates, degrees, and trainings (23.2%), and a network serving several science policy groups (1.0%). In contrast, at the national level, the preponderance of science policy programs are fellowships that place researchers in government (64.0%) or in government affairs offices (21.3%), followed by professional development and training opportunities (14.7%), and policy networks (2.7%). The growth in student organizations speaks to a rapid expansion of interest in science policy among early career researchers, driving the need for many other types of opportunities across the United States.

The nationwide science policy program database includes 13 from the Commonwealth of Virginia—academic certificates/degrees/training (6), student organizations (5), and government placements/fellowships (2)—the highest number in any state. Our data show through a pre-survey and interview with each program leader, that even though, at first glance, these programs reflect diverse models for ways to engage researchers in policy, they have many of the same goals for long-term impacts, similar conceptualizations of important audiences and activities, and shared philosophical orientations toward appropriate roles for scientists and engineers in policy and the nature of evidence use in policy. Furthermore, while these programs espouse evidence-based decision-making processes in governance, leaders find it challenging to identify tools that could allow them to implement long-lasting, effective approaches within their programs and to assess their impacts.

The long-term desired impacts Virginia’s science policy programs are seeking, based on our interviews with program leads, fall across four overarching domains: academic institutions, government policy processes, researchers' engagement in policy, and workforce development and training. While these programs may initially seem quite different based on audience, length and depth of program engagement, and type of activities undertaken by participants, they have much in common. When subdivided into 20 desired programmatic impacts, having “more scientists in policy” and facilitating “government employment of scientists” were most frequently cited. And leaders of all programmatic types—academic, student-run, and immersive government placements—seek to influence policy change as a long-term program impact.

Significance These findings suggest possible opportunities to scale and disseminate promising practices within this rapidly growing ecosystem. Supporting evidence-based approaches could both maximize programmatic effectiveness and advance our knowledge of how science-policy interfaces change and adapt over time. New capacity could be marshaled by building a broader network for shared learning between different types of programs and through partnerships between program leads and social scientists with expertise in the use of research evidence, policy processes, evaluation, and education.

ABSTRACT. Economists have long viewed innovation as the key driver of long-term economic growth. However, recent research has shown that discrimination and systemic racism have throttled the country’s growth path. This project seeks to evaluate how inequitable access to education limits full participation in the innovation pipeline.

Recent literature across the social sciences has consistently found that systemic factors limit who has the opportunity to innovate and that these disparities have quantitatively large detriments to society. Bell et al. (2017/2019) were the first to show, using linked taxpayer-inventor data, that children from lower-earning families, women, and racially disadvantaged groups had meaningfully lower innovation rates.

Given careers in innovation are often viewed as having risky payoffs that come later in life, we seek to uncover whether bigger loan burdens play a role in pushing talented students away from career paths as inventors. On average, Black students hold $7,000 more student loans than White students when they graduate, and this gap grows over time as repayment schedules diverge. Women hold 66 percent of the total outstanding student loan debt.

To better understand sources of disparities in innovation, we undertake a large-scale record linkage effort of national individual-level data on inventors from administrative records of the US Patent Office to student loan data from credit bureau records. The linkage allowed us to view 20 years of complete credit histories for almost 250,000 US patent inventors that have ever lived in California (about ⅓ of all US inventors). Applying imputation methods for race and gender, we find that 88% of our linked inventor sample is White or Asian (relative to 60% of the general CA population) and 17.5% are women.

Research is proceeding in two phases. In the first phase, which will be the primary focus of this talk, we document a number of stylized facts about credit histories and student loan usage among inventors. For instance, people with credit scores in the top 10% of the distribution are ten times as likely to invent as those with below-median credit scores. Inventors take out more loans of all types than the average person of a similar age. During their early 20s, inventors accumulate substantially more student loan debt than most people, but also pay off the debt more quickly; by age 40, inventors actually hold substantially less student loan debt than their peers. Contrary to what one might expect, event studies show that credit scores increase until the time of patenting, and then somewhat level off after patenting.

These descriptive results will lay the foundation for the second, more causal-oriented stage of research. We will test whether there is a causal relationship between education access and later-life innovation, and whether this effect can explain later-life disparities in innovation by race. We will use cutting-edge quasi-experimental techniques leveraging a series of policy changes to causally identify the effect of student loans burdens on later life innovation. Estimates will inform our understanding of the sources of disparities in innovation.

ABSTRACT. This work situates in the field of Science, Technology and Innovation Policy Studies, and its role to advance into sustainable human development processes (Arocena and Sutz 2003, Arocena and Sutz 2009, Sutz 2010, Alzugaray, Mederos et al. 2011, Arocena and Sutz 2012, Cozzens and Sutz 2012, Arocena 2019, Sutz and Bértola 2020). The study of science, technology and innovation is permeated by deep tensions regarding development dynamics, models of economic growth, production and consumption that today define western societies. At the same time, the current context poses dilemmas and challenges in multiple levels, where planetary instability is on the rise (UNDP 2020). Global wealth is increasingly unequal (Chancel, Piketty et al. 2021), strong unsustainable patterns perpetuate while other social and not as visible pandemias remain, as child labor, political polarization, and profound challenges that prevent the preservation of fundamental human rights (Giuliani 2018). References to civilization crisis, turning point (Biggeri, Clark et al. 2019), Anthropocene risk (Keys, Galaz et al. 2019), era transformation (Naciones Unidas 2016) point in the direction of the urgency for the development paradigm towards sustainability and social inclusion, with a long term approach. It is not clear how to process these transformations that challenge not only everyday life, but our institutions, principles and values, ethical imperatives, regarding human relationships but not only, also affecting non-human and environmental relationships in the path towards life forms that operate within safe planetary boundaries (Rockström, Steffen et al. 2009).

Even if the STI policy has gained an explicit space within the Latin American policy concert (Dutrenit, Aguirre-Bastos et al. 2021), the legitimacy is clear at the discursive level, but its intertwinement and embeddedness into crucial development aspects is still missing (Dutrenit and Sutz 2013, Bianchi, Bortagaray et al. 2021, Bortagaray and Aguirre-Bastos 2021). Furthermore, it is not evident what is best governance scheme to better advance in this direction, to foster the necessary changes across the policy spectrum placing STI as a fundamental driver of change. What type of policy space should STI have within the overall policy spectrum? Should it be organized as a discrete policy domain, or should it be designed in a way that cuts across others like agriculture, health, industry, social development? What scope, governance models and systems better serve such transformation, and with what specific policies, institutional arrays, and the extent to which they need to be changed or created? (Sachs, Schmidt-Traub et al. 2019). Current societal challenges are profoundly complex and novel, calling for flexibility, adaptation, experimentation and policy learning. STIP key (new) role is part of a “normative turn” (Daimer, Hufnagl et al. 2012), a paradigm shift in STI policy with growing importance of directionality and normativity in STI, and an instrumental role of STIP for solving societal challenges, and advancing sustainable human development.

This work aims at contributing to the overall discussion of STI policy with a specific focus on climate change. Even though the study has a national scope, the discussion and insights could be of international interest given the global character of climate change. More specifically, it has a twofold objective. First, the goal is to discuss a methodological approach to identify and assess the national research capabilities to then analyze the academic and policy needs that should be considered if tackling the national agenda on climate change. Secondly, and based on the study of the capabilities on climate change research in the country and the emerging opportunities, this work discusses a series of areas and knowledge gaps when it comes to advancing the decision-making processes on climate change in the country. For the identification of capabilities and opportunities related to climate change S&T, it is necessary to start from a perspective of complexity, which prioritizes the relevance of knowledge and innovation “for mitigation and adaptation to climate change and variability, as well as to reduce and adequately integrate uncertainties in decision making related to current impacts and future risks of climate variability and change” (Sistema Nacional Ambiental, Gabinete Nacional Ambiental et al. 2017). The work is based on different sources and types of information to account for the complexity that characterizes the identification of capabilities, opportunities and gaps in knowledge and innovation in climate change. Secondary information gathered from policy documents including plans and reports, complemented with data from a list of research proposals on climate change funded by the National Agency of Research and Innovation (ANII), and the University of the Republic between 2012 and 2021. Furthermore, the empirical analysis also relies on academic publications on climate change by authors whose organizational affiliation is located in Uruguay, from Scopus and Web of Science databases. Primary data gathered from in-depth interviews to qualified informants (first semester of 2022) is complemented with information from a survey to decision-making referents through an electronic form (second semester 2022). In summary, this work seeks to map the knowledge and policy needs related to climate change in Uruguay, understand the existing knowledge and innovation capabilities on climate change in the country, in order to identify the major dimensions around knowledge and innovation gaps in this area, to better grasp the potential role of the STI policy in relation to climate change, and the multidimensional and complex character of current development challenges.

ABSTRACT. This work examines the influence of different economic rationales in the Dominican Republic's (DR) science, technology, and innovation (STI) policies from a development context perspective, focusing on the design phase of the current Dominican STI policy framework. This work focuses on the influence of economic rationales to understand the formal construction of the contemporary Dominican STI policy platform, considering different economic policy cycles as contextual factors in creating an STI policy discourse as an element of development policies in one of the most dynamic developing economies of Latin America and the Caribbean. For this purpose, four STI policy frameworks are reviewed: the National Competitiveness Plan (PNC), presented in 2007; the Strategic Plan of Science, Technology, and Innovation 2008–18 (PESCYT), introduced in the fall of 2008; the Ten-Year Plan of Higher Education 2008–18 (PDES), presented in 2008; and the National Development Strategy 2030 (NDS), introduced in 2012.

The following questions are raised regarding these frameworks: What are the main rationales of the Dominican's STI policies? To what extent have the STI policies influenced this period? Given that four specific STI policy frameworks are considered in this work, it is necessary to highlight that STI policies generally do not operate independently. In its formulation, execution, and performance interact with and are affected by other policy frameworks. Limiting the current analysis to the four indicated STI policy frameworks is a decision to simplify the analysis and narrow it down in terms of its methodological scope. While recognizing the need for more complex analyses that deepen the structural, functional, and systemic interdependence of STI policies as part of the inherent complexity of innovation systems, it is imperative in the context of Latin American and Caribbean (LAC) countries.

Considering the LAC context, this work covers three cycles of STI policies in the DR. The first is the industrialization and import substitution cycle, corresponding to the development policies from the late 1960s to the late 1980s. The second one is the structural adjustment cycle, which is related to the regional influence of the Washington Consensus in the 1990s. Finally, the third one is the post-structural adjustment cycle from the twenty-first century's first decade. The main interest of this article is the third cycle since it was the period in which the current DR's STI policy frameworks emerged and gained political space.

From the methodological perspective, the identification of the rationales of the four frameworks of the STI policies of the DR started from a qualitative approach based on ground theory. Generally, grounded theory consists of an inductive method to construct a theory based on systematic review and data collection. Once the data are collected, a systematic review tends to support the discovery of theories based on the data. It builds a general description, which simultaneously fits with the theory and the data from which it emerges. The grounded theory approach tends to start with a couple of research questions about the phenomena under study, subject to an intense interaction between the data-gathering process and the analysis, making this approach suitable for the context.

Concerning the first research question raised in this article, and within the limits of the grounded approach in the Dominican STI policy frameworks, three rationales prevail: (1) a systemic–institutional nature, (2) a neo-Marshallian approach, and (3) a close-up Schumpeterian growth perspective. This general composition somewhat recognizes the systemic and institutional failures that still affect the growth and development potential of the country. However, it seems that this negative cycle can be broken through concerted policies, transparency, and considering a medium- to long-term STI policy formulation perspective.

Concerning the extent of the STI policy's influence on the Dominican economy's performance in the analyzed period, and without a proper evaluation of the impact of STI policies, it is not straightforward to determine the extent to which they contribute to economic growth and performance dynamics. Various factors, such as a favorable national and international macroeconomic environment, an increase in foreign direct investment, economic growth driven by aggregate domestic demand, tourism growth, and an increase in exports, can undoubtedly explain the prolonged growth cycle with greater robustness. However, there is no doubt that STI policies play at least a cultural role in defining a new approach to development policies in the DR. One possible hypothesis may be that STI policy frameworks have a latent effect (e.g., cultural, political, and institutional) when boosting and supporting economic growth in combination with other structural and cyclical factors. Another hypothesis could have a more direct effect, primarily through instruments and initiatives derived from the PNC, such as the neo-Marshallian influence in supporting SMEs and clusters through the creation of 'centropymes.' This is an interesting attempt to create nationwide SME–university–government interfaces. Such hypotheses, however, must be corroborated with further analysis.

In sum, rationales are not necessarily inserted into a specific policy framework without an explicit intention from the agents involved in STI policymaking. On certain occasions, this intention may be explicit, but it depends on the country's context and the interactions of the public policy agents.

The complex multilevel STI policy system built during the period covered in this article cannot be adequately assessed without creating an evaluation mix and conducting exercises, which are far from the purpose of this article. However, what can be achieved are more meditative, prospective, and general exercises regarding the possible contributions of the STI policies to the country's growth performance. In this regard, understanding the rationales of the Dominican STI policies sheds light on how to develop this evaluation mix, define the next generation of STI policies, and align them with sustainable development goals to boost productivity and competitiveness.

In terms of challenges and opportunities, there are at least four elements for the next generation of STI policies in the DR that could be briefly commented on: (1) overcoming coordination failures and improving governance; (2) rethinking business–university relationships; (3) incorporating an STI policy demand approach; and (4) applying a regional and multilevel approach to innovation systems.

ABSTRACT. Caribbean Small Independent Developing States (SIDs) face enduring economic challenges, including limited resources, geographic isolation, and structural vulnerabilities that constrain sustained growth and development. This paper examines the potential of Science, Technology, and Innovation (STI) policy as a mechanism to promote socio-economic transformation in Caribbean SIDs. Building on recent advancements in innovation policy and systems theory (Audretsch, 2007; Malerba, 2004; Seto et al., 2016), this study aims to address a pressing research gap: How can STI policies be strategically designed and implemented to drive economic transformation in the Caribbean SIDs context? The research is motivated by the critical need for Caribbean SIDs to diversify their economies and enhance resilience amid ongoing global economic and environmental pressures (Suárez & Erbes, 2021; Vargas & Darrow, 2023). Recognizing a gap in existing literature on STI policies tailored specifically for small economies, this paper adopts a multi-method approach that includes a systematic review of current STI policy frameworks, data analysis from Caribbean case studies, and comparative analysis of innovation systems in similar contexts. The study employs Systems of Innovation (SI) theory (Edquist, 1997), specifically exploring sectoral systems of innovation(SSI; Malerba, 2004), national systems of innovation (NSI; Nelson (1993), Lundvall B (1992), Freeman C. (1981), and regional systems of innovation (RSI; Cooke, 2001), to develop a multi-level policy model that addresses structural challenges faced by SIDs. This paper contributes to the literature by offering a policy framework tailored to the institutional and economic constraints characteristic of Caribbean SIDs, where traditional STI approaches may not readily apply. The model provides detailed policy recommendations designed to mitigate capacity limitations, institutional fragmentation, and sectoral dependencies. Furthermore, the study advances a SI-based model that facilitates scientific and technological integration within small economies, promoting pathways for sustainable economic growth and resilience (Gómez-Valenzuela, 2020; Weber & Rohracher, 2012). Data to support proposed models will be supported by secondary data sources and robust literature review on developing economies. In conclusion, this research underscores the importance of adaptive, innovation-driven policies for Caribbean SIDs and provides a strategic framework for advancing long-term resilience, economic stability and competitiveness in a rapidly changing global landscape. This framework addresses the specific needs of SIDs and contributes to broader STI policy discourse by emphasizing the potential of innovation systems in fostering socio-economic resilience in resource-constrained economies.

JEL: O31, O33, O38, R58, F63

ABSTRACT. <Background> China's rapid growth in technological competitiveness and its growing role in global commerce have left the US with concerns over its diminishing global influence. The US now seeks to exclude China from its high-tech supply chains, including semiconductor and EV supply chains, and improve the competitiveness of American industry. As the competition for technological hegemony between the US and China intensifies, the role of science and technology expands. Competition between the US and China has intensified due to the strengthening of US containment following the rise of China and China’s response to it. The evolution of major US policies to contain China, including trade sanctions, technological sanctions, to sanctions on the high-tech industrial ecosystem, confirms the US’ dependence on semiconductors abroad and the uncertainty of techno-politics. With the advent of economic security as the economy, security, and technology converge, economic security issues are becoming more important in science and technology areas such as climate change/energy, space/marine, and security of technology. In line with China’s economic rise, the US pursues new forms of economic security policies, such as trade policy, reorganization of global value chain, and strengthening of industrial policy. Security issues are involved in the science and technology sector, such as dual-use characteristics of space technology, weaponization of resources such as secondary batteries, supply chain issues such as carbon border tax in the field of climate change, and expansion of the scope of technology regulation due to the security of technology. In this situation, the suggestion of definition of SD and the directions for Korea's SD is requested to the SD center of KISTEP, which is designated by Ministry of Science and ICT and this research is the results of this request.

<Research Framework> Over the past 15 years, the definition of science and technology diplomacy has been discussed as a strategic and multidimensional concept and phenomenon with a clearer diplomatic purpose different from general international cooperation in science and technology, but there is still no consensus on the concept. Summarizing various discussions on the concept of S&T diplomacy, it is commonly defined as international partnership to solve common problems faced by mankind, such as global challenges, and recently discussed in consideration of international politics. Korea’s science and technology diplomacy is active in the type of ‘diplomacy for science’ (international cooperation in science and technology) and the perceptions of S&T diplomacy in the S&T group and diplomacy group are generally similar, but there are differences in terms of goals and important science and technology areas. Korea needs to establish the concept of science and technology diplomacy as it is recognized and commonly used as international cooperation from the perspective of science and technology and public diplomacy in the field of science and technology rather than the expression of science and technology diplomacy. Considering economic security, science and technology diplomacy is not only an area that utilizes scientific advice, but also an area that responds to technology control and secures a stable supply chain that considers economic and technology security (the latter is presented as a type of ‘diplomacy in science’). Forming in S&T diplomacy are areas, such as the area of climate change, where scientific advice is used in diplomacy, and where efforts are made to acquire stable supply chains and respond to technology control in consideration of economic security and technological security. The new convergence area of S&T diplomacy related to economic security and technological security is suggested as diplomacy in science, a new type of S&T diplomacy. Diplomacy in science counts as activities related to the control S&T for the securing of economic and technological security, supply chains and other desirables. The type of S&T diplomacy that aims for S&T purpose but employs such imperative means as the control of S&T for economic and technological security can be defined as diplomacy in science. While diplomacy for science has been rooted in cooperation and competition, diplomacy in science involves the acquisition of technology sovereignty and stable supply chains through standardization and control. Therefore, science and technology diplomacy can be classified into four types based on the policy goals (S&T or diplomacy) and the coercion of means (persuasive or commanding), and the type of ‘diplomacy in science’ is defined as a form of using commanding diplomatic means for science and technology.

<Results and Conclusions> Diplomacy for Science corresponds to the existing S&T international cooperation activities, including national strategic technologies, requires a more systematic cooperation strategy, and requires a bilateral/multilateral international joint research program led by Korea. Science for Diplomacy as science and technology diplomacy approached from the level of public diplomacy, it is necessary to promote expertise in science and technology in connection with the Indo-Pacific Strategy. Science in Diplomacy is a type of multilateral-oriented cooperation, such as solving global problems such as climate change and energy and achieving the UN’s SDGs and it necessary to strengthen its role by leading agendas suitable for its economic scale and global status. Diplomacy in science is a newly defined type of multilateral-oriented cooperation that requires strengthening cooperation with friendly bloc and continuing cooperation with the other bloc under rules-based regulation to secure strategic materials, rare resources, and stable supply chains. While the role of S&T diplomacy is expanding due to the US-China competition for technological hegemony, Korea is able to make strategic choices by virtue of its strength in semiconductors, batteries and other areas. The US-China competition for technological hegemony will likely continue into the future, and the role of diplomacy in science which supports economic security will likely expand.

ABSTRACT. What are the methodological approaches to studying innovation ecosystems across different disciplines? Is the innovation ecosystem becoming a unifying framework across disciplines, or is it divergent as research perspectives diversify? This study explores whether the innovation ecosystem concept converges into a unified theoretical framework or diverges across disciplines. To answer this question, we systematically reviewed 130 articles indexed in the Web of Science database, analyzing articles published over the past decade to identify core trends, theoretical foundations, and methodological approaches. A total of 130 references were identified, we traced 14,439 forward citations. Surprisingly, only one paper among the 14,439 forward citations was found to reference at least one paper from each of the identified thematic areas, highlighting the divergence in the existing literature. Thematic analysis of 130 papers sampled from this group revealed further methodological divergence and diverse research foci between network complexity, complementary forces, and ecosystem performance. Scholars interpret the concept to fit their disciplinary and methodological training rather than engage in cross-disciplinary research.

ABSTRACT. The shifting geopolitical landscape is creating new challenges for small economies in Asia. At the same time, the competition among major powers adds further complexity to their efforts in managing technology innovation. However, limited scholarly attention has been given to how geopolitics shapes the biotech sector in these contexts. This study addresses this research gap by conducting a comparative case study on Taiwan and Singapore, exploring the impact of geopolitics on biotech innovation in these two Asian economies. Through a mixed-methods approach that includes primary interviews and secondary document analysis, this research identifies three key factors influencing biotech innovation policies and outcomes: risk perception, international institutions, and state-capital relations. The findings show that for small economies like Taiwan and Singapore, geopolitics has become a critical driver in shaping innovation policies and fostering state-firm collaboration. However, differences in how these economies perceive geopolitical risks, nurture international institutional environments, and manage relations between the government and domestic as well as multinational capital have led to divergent strategies and outcomes in biotech innovation. In Taiwan, the heightened perception of geopolitical risk, particularly in light of cross-strait tensions, has driven a policy focus on self-reliance in biotech capabilities. The Taiwanese government prioritizes domestic talent development and technology to enhance national resilience. On the other hand, Singapore adopts a proactive and outward-looking strategy, leveraging its geopolitical stability to attract foreign investments and expertise. By positioning itself as a global biotech hub, Singapore emphasizes strong collaboration with international institutions and multinational corporations. This study offers an innovative analytical framework for understanding how small economies navigate the geopolitics of emerging technologies in high-risk and capital-intensive sectors like biotechnology. It highlights how the interplay of geopolitical risk perception, institutional support, and state-capital dynamics can drive distinct policy choices and innovation trajectories. The research offers practical insights for policymakers in small economies on optimizing innovation policies. It highlights ways to balance national interests with global collaboration to foster sustainable growth in the biotech sector.

ABSTRACT. The medical device industry is rapidly growing, with great potential to improve healthcare outcomes worldwide. Taiwan and India have been focusing on building their medical device innovation systems to enhance competitiveness. According to the Multi-Level Perspective (MLP), transitions occur through interactions among landscape, regime, and niche. While specifics vary, niche innovations develop internal strength, and landscape changes exert pressure on the system, creating opportunities for disruption (Geels, 2002; Kemp, Rip & Schot, 2001; Kemp, Schot & Hoogma, 1998; Schot et al., 1994). How do institutional factors facilitate the transformation of the medical device innovation system in Taiwan and India? This research adopts the national biotechnology sectoral innovation system as the main framework. The study adopts a mixed-method approach, combining the in-depth interview and system dynamics. The results show that Taiwan focuses on the development of high-tech medical devices and India on low-cost medical devices. In Taiwan, the medical device industry utilizes technological niches to create opportunities for emerging technologies and interdisciplinary integration, fostering the establishment of actor networks and contributing to transformative changes in the landscape. In India, the government has supported the medical device industry through initiatives such as the "Make in India" campaign and the establishment of the medical devices park to provide a conducive environment. Overall, this paper provides insights into the strategies and initiatives that Taiwan and India have implemented to build their medical device innovation systems and offers recommendations for policymakers and industry leaders further to enhance their competitiveness in the global medical device market.

ABSTRACT. How does the high-tech sector in Taiwan deepen integration within the global innovation ecosystem? How do science and innovation policies strengthen international collaboration to foster sustainable economic and social impact? Artificial Intelligence (AI) has become a pivotal force shaping both economic growth and socio-technical transformations worldwide, especially in the context of rapidly evolving global trends and challenges. Since 2016, the United States, European Union, Canada, and other countries have enacted policies focused on AI research, ethical governance, and competitive strategy. Taiwan has also enacted a series of AI-related policy initiatives, including the Artificial Intelligence Development Basic Act (2019), the Artificial Intelligence Basic Law (2023), and the Artificial Intelligence Fundamental Act (2024). This study explores the AI innovation ecosystem in Taiwan, focusing on how it aligns with and contributes to global science and innovation policy trends, especially in the context of U.S.-China tensions and shifting geopolitical landscapes. This study adopts a mixed methods approach, including scientometrics, social network analysis (SNA), documentary analysis, and system dynamics, to explore the development of AI innovation ecosystem in Taiwan. The study finds that the AI innovation ecosystem in Taiwan is developing rapidly, demonstrating competitiveness in areas like hardware manufacturing and intelligent biomedicine. However, Taiwan faces challenges in enhancing its AI software development capabilities. This paper analyzes the system dynamics among regimes, technology niches, and the landscape of the artificial intelligence ecosystem. Additionally, this research will shed light on the systemic transformation of Taiwan’s technological soft power and economic competitiveness.

ABSTRACT. How does the innovation and entrepreneurship ecosystem evolve? What's the role that incubation centers play in the emerging entrepreneurship ecosystem? This research investigates the evolution of incubators and start-up accelerators, focusing on their mediating roles in fostering innovation and entrepreneurship while navigating complex policy and system dynamics. Since the late 1990s, Taiwan’s academic incubation centers have transitioned from entrepreneurial space providers to entrepreneurship platforms, bridging research outcomes and entrepreneurial ventures. The study examines three types of start-up accelerators: university-based incubation centers, research institute-based incubation centers, and enterprise-operates accelerators. Combining mixed-method approaches, including 20 in-depth interviews, documentary analysis, system dynamic analysis, and social network analysis, this research explores the transformation of startup accelerators and the mediating roles these knowledge brokers play in the emerging entrepreneurship ecosystem. The results show that Taiwan effectively promoted technological innovation and university-industry collaboration since 1998 by establishing academic incubation centers and technology transfer offices. However, after 2016, shifts in government funding priorities and policy changes forced the incubation centers to transform into start-up accelerators. These policy changes facilitated the dynamic evolution of the entrepreneurship ecosystem. We also found incubation centers alone are not sufficient for incubating high tech startups. In the science based sector, firms to play more effective brokerage roles may be a more successful strategy to develop a science-intensive sector. At the macro level, based on a multi-level perspective framework, this research provides insights to illustrate the mutual shaping of innovation policies and the development of the entrepreneurship ecosystem. The regime plays a central role in responding to landscape pressures, while niches serve as platforms for experimentation and innovation that carry out broader transformations. Ultimately, this research attempts to offer policy recommendations for speeding up knowledge transfer, enhancing academia-industry collaborations, and strengthening the innovation and entrepreneurship ecosystems.

ABSTRACT. The roundtable examines the critical, yet often underexplored, roles of "boundary-spanning" professionals—those who bridge the divide between policy development and implementation within innovation ecosystems. This group includes technology transfer specialists, evaluators, consultants, and project managers. The Boundary-Spanning Theory—originating from Tushman (1977) and applied in various policy and organizational studies—explains how these professionals facilitate interactions and knowledge exchange across organizational and cultural divides, thereby driving the effective translation of policy into practice.

In the context of innovation policy, these roles are essential for implementing public-private partnerships, disseminating research, and managing complex, cross-sector projects. Boundary-spanners act as conduits for both technical expertise and institutional knowledge, ensuring that innovation policies achieve their intended impacts. This panel will discuss:

- How these boundary-spanning roles have professionalized within the policy life cycle? - The distinct skills and competencies that are crucial for success in these roles? - The ways that boundary-spanners contribute to a more agile, responsive policy implementation process?

Additionally, the panel will explore the broader implications of these roles for building a resilient and adaptive innovation ecosystem. The need for boundary-spanners reflects an underlying shift in how innovation policies are conceived and executed, emphasizing collaboration, accountability, and inclusivity. Panelists will discuss how organizations can foster environments that empower these professionals, such as creating clearer pathways for career advancement and providing resources for ongoing skills development.

The session will draw insights from case studies across academia, government, and industry, illustrating how boundary-spanners have contributed to successful policy implementation in diverse contexts. For example, technology transfer offices in universities often bridge the gap between academic research and market applications, while evaluators and consultants ensure that policies meet their intended goals by providing rigorous assessments and recommendations.

By examining both the organizational and human dimensions of boundary-spanning roles, this panel will provide actionable insights into how the workforce can adapt to meet the challenges of modern innovation policy. Attendees will leave with a deeper understanding of how these professionals drive the success of complex initiatives and how investments in their training and development can strengthen the broader innovation ecosystem.

ABSTRACT. Women are underrepresented in science across most countries, particularly in senior roles. Our study investigates one of the possible factors contributing to the persistent gender differences in top academic and scientific positions, namely bias in the evaluation for promotion to full professor and senior researcher. We study gender differences in promotion probabilities at the Spanish National Research Council (CSIC) in a five-year period from 2017 to 2021. We analyze 90 different promotion events evaluated by singular/independent evaluation panels, across scientific fields. Those include 2338 applicants and 397 members of evaluation committees. We address three research questions: 1. Whether merit factors (and which ones) positively and significantly affect individual promotion probabilities. 2. Whether the gender of candidates is a significant factor when other relevant variables are considered. 3. Whether there are factors associated with the composition of the panels and the design of the promotion process that affect unequal probabilities for candidates.

Our study provides the opportunity to study promotion processes in an institutional setting characterized internal competition and based on a contest-type process, different to the one predominant in the US research university model most typical in a large part of the international literature, namely, the tenure-track model, where, moreover, hiring and promotion evaluation takes place at department level. In contrast, in the model analysed in this paper evaluation panels operate at the level of the position specialty and include external as well as internal members from several institutes. Understanding how competition levels are influenced by organizational decisions, such as the distribution of positions across scientific fields and categories, enhances our insights into how organizational practices impact individual promotion probabilities, especially in contested systems where the probability of promotion depends on the number and characteristics of competitors

We have created two different data sets; one of applicants, combining personal and career information with scientific performance data (publications, citations, impact, projects, contracts, patents, and PhD supervision) and another one of Evaluation Panel members. For the analysis, a unique database of the 2,358 candidates has been created, incorporating demographic, career and merit variables; a database of the members of the evaluation panels who have participated in the evaluation processes has also been built. These two databases have been linked through the joint occurrence of evaluators and applicants in the event of every contest for the promotion positions. We have integrated the two data sets for each of the 90 evaluation panels and generated additional indicators such as share of females’ evaluators in every panel and connections between applicants and Panel members. Some characteristics of the promotion processes (number of available positions and its allocation by field areas) condition the promotion probability of candidates, so calls corresponding to 5 years have been selected with different levels of success rates, for 2 senior scales, with 90 different scientific profiles (from different scientific areas) with evaluation committees, etc. Using logistic regression models and estimating predicted probabilities, we address the questions of weather gender plays a significant role in promotion probabilities, the role of merit, and the effect of competition and panel composition and connections with applicants.

We do not find consistent evidence of systematic gender biases in the promotion processes. Evidence shows that female and male applicants possess comparable qualifications and merits. While we find that gender is not relevant to explain individual promotion outcomes, scientific achievements (specially research funding and impact) are important predictors of promotion probabilities across genders. The merit factors introduced in the model have positive effects (most of them), although some are not significant. Of all the merit factors, it turns out that the most important factor for both promotion categories is, firstly, " being the IP of research projects" either considering the number of projects and/or volume of funding (ln), and secondly, the citations (fractional count) received by candidates. Our findings reveal additional influential elements beyond scientific merits. Specifically, institutional proximity between applicants and evaluators is a relevant significant factor, in addition to the effects of competitive pressure within each promotion event, however those effects are not gendered.

We find a powerful and regular effect in favor of the probabilities of candidates who have an evaluating member from their own institute in the panel. Having a member from the own institute in the evaluation panel (a measure of institutional proximity) significantly increases promotion probabilities (for both sexes), by almost 50% compared to those with similar characteristics but without that evaluator from the same institute.

Another interesting finding is that when there is more than one candidate from the same institute, their promotion probabilities are reduced, limiting the positive effect of having a colleague of the same institute on the panel.

Our Contribution First, our study is not sample, it provides an analysis of all promotion processes and therefore covers different fields. Few studies have previously simultaneously addressed a variety of scientific fields. Second, while other studies have focused on specific phases of academic careers, with a predominance of access to tenure, we address promotions at two different senior scales, corresponding to different career phases, offering complementary insights into the factors that influence progress to the highest academic ranks. Third, we have diversified the indicators of the merit of applicants, expanding beyond bibliometric ones. Fourth, the analysis focuses on a large research institution (CSIC), with scientific and technological research as the main mission, avoiding measurement problems associated with teaching loads. Fifth, although recent studies already included the analysis of evaluation panels (especially referring to funding agencies), the incorporation of relevant attributes of evaluation panels’ members is not yet not common in promotion studies. Finally, from the policy side, some of our results raise concerns about the fairness of the promotion process. We discuss the findings and some organizational policy implications, highlighting the importance of examining not only individual level factors but also process and structural ones.

ABSTRACT. Fighting inequality in the research ecosystem has become an increasingly prominent topic of discussion among scientists, universities, editorial bodies, and research institutions. The various inequalities faced by scientists significantly impact their activities, career trajectories, and workforce development. Particular attention should be given to the crucial role of funding agencies (FAs) in addressing inequalities in the scientific environment. As key players in the research ecosystem, FAs possess the power to either drive institutional change or reinforce, perpetuate, and even create new forms of inequality. Although there is a concern on the part of these institutions to promote a diverse and inclusive environment, it is still a challenge to: i) identify the need for interventions and categorize them; ii) determine their efficacy and effectiveness; and iii) identify what characterizes the institutional capacity of an organization to determine why certain initiatives are effective and why some of them fail in promoting Equity, Diversity and Inclusion (EDI) in the research environment. This study developed an original framework to map the range of EDI activities across various contexts and identity characteristics of FAs. It also aims to examine the institutional capacity of FAs to effectively implement EDI initiatives. The guiding research question is: How are EDI-related initiatives being institutionalized by FAs? We conducted a literature review on the topics of EDI data and metrics to provide an updated perspective on how this issue is being addressed in recent research. Based on the review and benchmarking of previous studies, we developed seven categories of analysis to evaluate FAs institutional capacity to address inequities in science. Our preliminary results draw on secondary data from the selected FAs, including an analysis of their recent strategic plans, gender-related reports, and initiatives addressing EDI-related elements. The next step involves conducting interviews to collect primary data. We aim to compare the initiatives, common approaches, and innovative activities of the selected funding agencies.

Methodology

The research aims to identify initiatives that incorporate elements of EDI within research FAs. To achieve this, a framework was developed to collect secondary data from FAs, utilizing official documents such as strategic plans, annual reports, EDI-related action plans, and statements, as well as information available on official websites. The first step of the methodological approach involved conducting a benchmarking and literature review to create categories for EDI-focused initiatives. These categories will subsequently be used to measure the institutional capacity of the analyzed FAs. The data sources used for this initial stage included scientific articles and reports that address the issue of inequality in science and on the collection and monitoring of diversity data. To measure the institutional capacities of organizations to develop and implement effective EDI strategies and interventions, we investigated the following aspects and classified them into implemented practice, being implemented, and not available information. The seven categories, as well as the capabilities and initiatives evaluated, are: i) To mention EDI in official documents ii) Axes of diversity (gender, race, ethnicity, socio-economic, disabilities) iii) To have dedicated EDI departments, teams, or other sorts of institutional areas iv) Operationalization (e.g.: to have policy or institutional guidelines on EDI directed to the FAs' staff, reviewers and grantees; to have specific calls for underrepresented groups aimed at promoting EDI…) v) To have internal training on EDI concepts, practices and policies for staff, reviewers, applicants and grantees vi) Data collection of EDI of applicants, grantees, reviewers, and decision boards vii) To evaluate and disseminate (to monitor and measure progress systematically of EDI initiatives; to make aggregated data publicly available; to communicate to the scientific Community and to the general public) We consider all categories at the same level of relevance. Therefore, we assign equal weight to each category in the institutional capacity assessment, with each category receiving a maximum score of 10 points. The following step is to apply this framework to selected Funding Agencies. Our study considers the application of this framework with two national FAs in Brazil, the Brazilian National Council for Scientific and Technological Development (CNPq) and the Coordination of Superior Level Staff Improvement (Capes); the National Council of State Foundations for Research Support, which currently brings together 27 FAs in Brazil and to a private FA called Instituto Serrapilheira.

Preliminary findings

This study is a work in progress, and we present here only preliminary findings based on secondary data from the analyzed funding agencies FAs. Despite the growing attention to EDI, there appears to be a predominant focus on gender-related initiatives, as exemplified by programs from both CNPq and Capes. In contrast, an intersectional approach that considers additional diversity dimensions, such as race, ethnicity, and disabilities, is more evident in activities promoted by the Instituto Serrapilheira. The use of specific calls targeting underrepresented groups to promote EDI is more common among agencies than the inclusion of EDI criteria in general calls. Another important point to highlight is the issue of data availability. The decision not to communicate or demonstrate a specific program or initiative, or the lack of clear information about a strategic plan, for example, may be interpreted as a lack of maturity concerning the theme of combating inequality, or can indicate that the implementation of EDI can face more institutional obstacles and advance in a slower path. Instances of this nature were observed in The São Paulo Research Foundation (FAPESP), often leading to a need for more available information and reliance on news reports from the agencies rather than official documents published by them. The absence of readily available information or institutionalized documents, in some way facilitated the identification of areas and practices resistant to change. This should be further investigated and confirmed with the gathering of primary data through interviews and questionnaires with the selected FAs. Investigating how FAs, a core element of the research system, are addressing this issue can be an important starting point to assess if real change is actually taking place.

ABSTRACT. In 2018, we delivered a survey to scholarly parents to understand the gendered relationship between parenting and productivity (Derrick, et al., 2022). Our responses, received from 10,445 active scientists around the globe showed a strong parenting penalty based on levels of parenting engagement. That is, while gender differences were observed, they were largely driven by the higher engagement of women in parental duties. In 2020, the onset of the social restrictions wrought by the pandemic had a demonstrable immediate effect of production of science by younger women globally (Vincent-Lamarre et al., 2020). Universities responded by creating policies around COVID: allowing for extensions to evaluative periods (e.g., tenure-clock extensions in the United States) and introducing “COVID impact statements” into evaluations (Malisch, 2020). Surveys conducted nine-months after the pandemic suggested that there may be longer term production consequences, particularly for women with children (Gao et al., 2021) and many of these accommodations are still in place, indicating that the effects of the pandemic may be experienced long after social restrictions have ceased. This study seeks to examine the long-term consequences caused by COVID on academic productivity for parents considering not only the disruption of COVID by the potential recovery period to return to pre-COVID levels of productivity.

Using authors in Web of Science as a population, we constructed a sampling frame of one million disambiguated authors associated with a unique email address who had published at least one work as a corresponding author within the last five years and at least three papers in their publication history. The survey replicated the questions from the 2018 survey, but added additional questions focused explicitly on the experience of parents during the pandemic. This resulted in data that examined three time periods: pre-pandemic (2018); pandemic (2020-2021), and post-pandemic (2024). We received 7291 completed surveys at the time of this analysis. These responses were then matched back to productivity data within the Web of Science to compare parenting engagement with productivity.

We analyzed the average number of publications per year by the cohort from 2000 to 2022 and found a relatively stable increase, peaking, for both men and in 2019 and experiencing decline since 2022, suggesting a productivity loss that has not recovered (Figure 1). We also see that this sample contains several women with higher rates of production in the earlier years and a decline in the later years. This is a function of age: when we limit the population to only those who published their first paper after 1992, we find a much stronger rate of increased productivity and sharper rate of decline (Figure 2). The filter question for the survey was whether the respondent had children. Several individuals started the survey but then indicated that they did not have children. These data are mapped as “no kids” on Figures 1 and 2. As shown, there is no clear relation for the younger cohort between those with and without children, in terms of productivity.

We then relate these data to parenting engagement, to understand the effect of parenting on rates of production. We examine the difference between the number of papers published in 2020-2022 and those in 2017-2019. As shown in Figure 3, those individuals who had children prior to the pandemic experience a dramatic loss in productivity as compared to those who had their children in 2020 and beyond. We also see that men had a greater loss of productivity during the pandemic, compared to women.

Given these observed differences, we examine whether degrees of parenting engagement also varied during this period. As shown (Figure 5), women shifted strongly into lead parenting positions during the pandemic, sharply decreasing in satellite parenting and moderately declining in dual parenting. Men had a large increase in dual parenting, mirrored by a decline in satellite parenting. Overall, levels of parenting engagement increased for all parents during the pandemic and particularly for men. Post-pandemic, however, women shifted back to dual parenting, while men rebounded away from lead and dual to more satellite parenting. The return, however, was not back to 2018 levels for either—showing higher levels of post-pandemic parenting engagement.

Differences, however, varied dramatically by country. For example, while the United States (which comprises about a third of all respondents) mirrors the global level, variations are observed for other countries. Women in the UK, e.g., had much stronger lead roles during the pandemic and men maintained a slightly higher rate of dual parenting post-pandemic. Nordic countries observed an increase in women in lead positions during the pandemic, with women taking strong dual roles in the post-pandemic period.

Building on these initial results, we will explore the survey to gain a deeper understanding of the disruption of the pandemic on academic parents, whether this disruption was gendered, how it varies across country and discipline, and dynamics of the recovery period. The goal of this study is to inform policy makers to the extent of disruption and how policies can account not only for the immediate event, but the lingering consequences of this event on expected productivity, particularly across countries and by gender. This work also informs the broader body of literature on the post-pandemic labor market and sociological changes to work-life balance in the aftermath of COVID.

REFERENCES Gao, J., Ying, Y. Myers, K.R., Lakhani, K.R., & Wang, D. (2021). Potentially long-lasting effects of the pandemic on scientists. Nature Communications, 12, 6188. Malisch, J.L., et al (2020). In the wake of COVID-19, academia needs new solutions to ensure gender equity. PNAS, 117(27), 15378-15381. Vincent-Lamarre, Sugimoto, C.R., & Lariviere, V. (2020). The decline of women’s research production during the coronavirus pandemic. Derrick, G.E., Chen, P-Y., van Leeuwen, T. Lariviere, V, & Sugimoto, C.R. (2022). The relationship between parenting engagement and academic performance. Scientific Reports, 12, 22300.

ABSTRACT. Media plays a significant role in communicating scientific discoveries to the public and possesses the power to influence the public perception of science and science policies. The representation of diverse groups of scientists and their research is one of the basic concerns of equity in science journalism. Existing studies suggest systematic gender bias in media coverage of science can affect the citation, funding, collaborations, and career advancement opportunities for certain groups of scientists and influence young individuals' career choices. Such biased coverage of science based on gender not only hinders the communication of diverse scientific topics but also creates a biased view of who can be a scientist. Therefore, understanding and addressing gender bias in science coverage is essential for fostering a more informed and equitable public discourse on scientific issues. In this study, we intend to investigate the existence of gender bias in media coverage of scientific research.

This study primarily relies on data from altmetric.com (Altmetrics) and Scopus. We selected the top 25% of journals indexed by Altmetrics with the highest news mention number for each of the 20 fields specified by the Science-Metrix classification and collected the DOIs associated with those journals. For these DOIs, we collected the news citation data from Altmetrics. We limit our analysis to the 1,014,077 unique articles published between 2018 and 2022 within the United States.

To evaluate the impact of gender on media citations, we created a matched sample of articles that were not cited in the media (control group) but exhibited similar characteristics to those cited in the media (treatment group) using Coarsened Exact Matching (CEM). The characteristic list includes publication year, publication venue, subfield of the article, number of authors in the author list, and open access status. Our analysis reveals that in fields such as Historical Studies, Mathematics & Statistics, Biomedical Research, Social Science, and Public Health & Health Services, the proportion of women corresponding authors is notably lower in the treatment group. Conversely, we observe a higher proportion of women corresponding authors in fields such as Economics & Business, Philosophy & Theology, and Agriculture, Fisheries, and Forestry.

We conducted a weighted binary logistic regression analysis using the CEM weights. Our analysis suggests mixed results. While male corresponding author papers are significantly more likely to be cited in fields like Public Health & Health Services and Social Science, we see female corresponding author papers to be significantly more likely to be covered in Economics & Business, Information & Communication Technologies, Agriculture, Fisheries & Forestry, and Psychology & Cognitive Science. This pattern of media coverage draws an interesting connection between media citation and the gender equality paradox. In fields with a higher representation of women corresponding authors, papers by women are less likely to receive media citations. In contrast, fields with a lower share of women corresponding authors show an increased likelihood of media citation for papers authored by women.

We further compared the raw number of media citations received by papers with women versus men as corresponding authors, focusing only on papers with at least one media citation. Our analysis reveals a difference in the average number of media citations, favoring male corresponding authors across all five domains. We conducted a regression analysis to find the association between gender and raw media coverage count while controlling for year, team size, the subfield of the article, and journal quantile across domains. Our analysis suggests that male corresponding authored papers receive a significantly higher number of media citations across the majority of the domains.

We analyzed the correlation between the gender of the corresponding author and the likelihood of being cited across different media types: local, national, international, and specialty. Our findings suggest that the proportion of women corresponding authors is higher in the locally cited group than in the nationally, internationally, and specialty cited groups, indicating that women corresponding-authored papers receive more citations than the control group in local media, while other broader-reaching outlets and specialty media cite more men corresponding-authored papers. We also assess the impact of the gender of the corresponding author on the ideology of the media outlet. Our regression analysis suggests that papers by women corresponding authors are more frequently covered by outlets with a liberal ideological stance compared to papers by male authors for overall trend and in Natural Sciences, Health Sciences, and Economic & Social Sciences.

To estimate the impact of gender on different types of sentiment scores, we predicted sentence-level sentiment scores for each news article. Using the model, we predicted sentence-level positive, negative, and neutral sentiment scores. We conducted regression analysis at the sentence level, holding control variables constant. Our analysis reveals that papers with men corresponding authors received significantly higher positive sentiment scores in media coverage, both in the overall trend and specifically in Applied Sciences and Health Sciences. Additionally, we found that papers with men corresponding authors received significantly higher neutral sentiment scores overall. Conversely, papers with women corresponding authors were cited in media with significantly higher negative sentiment scores, with this trend holding across Applied Sciences, Health Sciences, and Natural Sciences. However, Arts & Humanities presented an exception: in this field, papers with men corresponding authors were associated with significantly higher negative sentiment scores. These findings indicate that media coverage sentiment may be influenced by the gender of the corresponding author, with women-authored papers being more often associated with negative sentiment, suggesting potential gender bias in media coverage of science.

Overall, these results point to systemic issues in the representation of scientists in media. The existing gender difference in media citations and sentiments may perpetuate gender disparities in scientific recognition and public influence. Addressing these biases is crucial for fostering equitable representation of scientists across genders. Media outlets and journalists should critically evaluate their reporting practices, ensuring that coverage fairly reflects the significance and quality of research regardless of the author’s gender.

ABSTRACT. This panel will examine factors relevant to new investments in regional place-based innovation initiatives.  Many of these programs target regions that have not traditionally been innovation hubs with the goal of expanding economic and other opportunities for these communities. These investments in "smart innovation"—a collaborative, equity-centered approach to innovation—are a strategy for reinforcing democratic foundations and fostering regional resilience. In the current landscape, characterized by socio-economic divisions and structural disparities, "smart innovation" provides a model that emphasizes shared benefits and cross-sectoral partnership to address regional needs and foster inclusivity. The panel will explore how frameworks like the Triple Helix Model, Knowledge Spillover Theory, Institutional Logics Theory, Boundary-Spanning Theory, and models of Co-Production reveal both challenges and pathways for realizing democratic goals within innovation policies.

The Triple Helix Model underscores the need for interconnectedness among academia, industry, and government, suggesting that balanced partnerships across these sectors are key to aligning innovation initiatives with democratic values and inclusive economic development. However, divergent institutional logics—where government regulations, academic objectives, and industry imperatives may clash—pose significant obstacles. Using Institutional Logics Theory as a backdrop, the panel will discuss strategies for mitigating these tensions to foster cohesive policy alignment that serves public interests.

Central to this discussion is also the role of knowledge spillovers, where technological and organizational insights flow between entities, generating broader economic and social benefits. Knowledge spillovers—first discussed in economic and innovation literature by Griliches (1992) and expanded upon by Link and others—are crucial for creating dynamic, resilient innovation ecosystems. The panel will examine how government bodies, academic institutions, NGOs, and private industry can collaboratively ensure that the benefits of innovation extend to regions that might otherwise be left behind.

Boundary-Spanning Theory further illuminates the importance of individuals and organizations that bridge these sectoral divides, ensuring cohesion between varied stakeholder goals. These "boundary spanners" facilitate knowledge exchange and collaboration across traditionally isolated groups, fostering a shared understanding that is essential for smart innovation to thrive. Finally, these initiatives often challenge individuals and organization to engage in knowledge co-production in order to meet regional needs. Co-production presents additional challenges in the dynamics of how partners from multiple sectors work together to advance regional opportunities and impacts.

Using examples from regional development projects like NSF programs, rebuilding due to conflict and natural disaster, and global innovation initiatives overall, we will discuss:

- How do we design inclusive policies that adapt to both local and national priorities?

- What are the key components of a sustainable, democratic innovation ecosystem?

- How can metrics be established to measure the societal impact of innovation, beyond traditional economic metrics?

Panelists will address these questions by examining different stakeholders’ roles in fostering knowledge-sharing and democratic accountability.

Altogether, this session positions "smart innovation" as a vehicle for resilience and democratic stability by embedding equity, accountability, and collaborative governance into innovation-driven development. The panel will engage participants in discussion on how these frameworks can serve as guides for both policy design and implementation in ways that strengthen democracy and build sustainable growth at multiple levels.

ABSTRACT. This paper will look at attempts in a selection of European Union member states, and the UK, to balance internationally collaborative research activities and a commitment to open science with the perceived need to consider issues of research security and ‘responsible internationalisation’ in a new era of economic and geopolitical competition. National security and economic security concerns have been a key driver of national research policies in most countries, but over many decades international mobility and collaboration have increased in importance, shaping the emergence of a globalized science system (Wagner, 2009). Governments have historically often attempted to restrict the flow of technological artefacts, codified knowledge and tacit know-how in order to maintain technological leadership, but today’s globalized science system and a renewed interest in ‘open science’ and open data makes the challenge of doing so much more complex.

European nations differ significantly in the level of internationalization of their national public research systems, in the institutional configuration of those systems, and in terms of how developed policies and practices are in relation to research security, but the issue of research security is rising rapidly on the policy agenda of all European countries and awareness, and good practice is increasingly being promoted through the action of international bodies such as the European Comission, OECD, and G7, and through organisations such as the League of European Research Universities (LERU). For instance, in January 2024, the European Commission Council Recommendation on Research Security highlighted a need to enhance research security across Europe and also announced the creation of a European centre of expertise on research security, and work on tools to support research performing organisations and reforms to the EU’s ‘dual-use’ export control regime. For its part, the UK has built a relatively sophisticated system of research security measures through close collaboration between government, security agencies, research funders and universities. However, there is little international comparative research on research security concerns and responses (Shih, Chubb and Cooney-O’Donoghue, 2024).

The term ‘research security’ is commonly used in, and well understood by research system actors in the UK and other Anglophone countries, but ideas around ‘trust’ and ‘responsibility’ also figure in the language that countries use to discuss these issues. For instance, the concept of ‘responsible internationalisation’ provides a more positive take that implies a recognition of the value and importance of continued internationalisation whilst de-emphasising the notion of restrictions on academic freedom to share and collaborate, whilst the concept of ‘trusted research’ links concerns about the protection of knowledge, know-how and IP with concerns about openness, integrity and reproducibility in research.

Based on interviews with research system actors in seven EU member states (Czech Republic; France; Germany; Italy; Netherlands; Spain; Sweden), and documentary analysis, we examine evolving understandings, regulations and practices in relation to research security in the public or quasi-public research base. The seven countries were selected based on two criteria: the importance of their publicly funded research base to the overall European science base; and coverage of a variety of national research systems configurations including one post-Soviet Academy of Sciences-style system. We also draw on interviews at the EU level and with informants from relevant networks of research-performing organisations.

We look at the way different systems are attempting to assess and mitigate perceived risks stemming from international cooperation and the role of science ministries, funding agencies and other key actors in this. We explore how the tensions between ideals around academic freedom (but also pre-existing cultures of practicing research, including research collaboration and sharing of data and knowledge) and new pressures to secure against perceived national and economic security risks of knowledge leakage or theft. We also explore the role of international organisations and networks in shaping perceptions and sharing practices.

The aim of the paper is to draw lessons from the comparative analysis about how the potential tensions between openness and security are perceived, are being managed, and how ‘trust’ in research is being (re)constructed in this new era of responsible internationalization.

Shih, T., Chubb, A. & Cooney-O’Donoghue, D. (2024) Scientific collaboration amid geopolitical tensions: a comparison of Sweden and Australia. Higher Education 87, 1339–1356. https://doi.org/10.1007/s10734-023-01066-0

Wagner, C. S. (2009). The new invisible college: Science for development. Rowman & Littlefield.

ABSTRACT. Abstract submitted as part of the proposed thematic session: "As open as possible, as secure as necessary": A Discussion of the Securitization of World Scientific Research and Development.

The challenge of maintaining international scientific collaboration while safeguarding national and economic interests has led governments to pay increasing attention to research security. Research security refers to a broad range of issues and risks associated with international research collaboration, including threats to national security and other risks such as attempts by external actors to illegitimately acquire academic research and expertise and/or interfere with academic discourse (G7, 2022).

This paper will examine the United Kingdom’s approach to research security, what the UK government calls “Trusted Research”. Specifically, it will examine these developments through the theoretical lens of collaborative governance, highlighting the tensions and trade-offs between the co-development of research security measures between government, researcher funders and universities and the threat of more adversarial relationships.

The issue of research security is particularly important for the UK since international research collaboration, not least with the People’s Republic of China, plays a key role in UK academic science. Studies suggest that virtually all the growth in UK science (as measured by publication) over the last two decades is accounted for by growth in international collaborative papers, including with China (Adams et al, 2022). At the same time, Chinese students account for around one-third of all UK PhD students in STEM subjects and UK universities collaborate with Chinese institutions in a substantial number of joint research centres (Quimbre et al, 2022).

The paper considers Trusted Research and its core principles. This places an emphasis on the responsibility of individual researchers and institutions in managing risks, promoting informed decision-making, and protecting research from hostile actors. The Trusted Research campaign, led by government but co-developed with university associations and research funding bodies, seeks to raise awareness of threats amongst universities and academic researchers, diffuse good practice with respect to due diligence in the examination of international research collaborations and provide guidance to researchers, university staff and funding organisations (Universities UK, 2022; NPSA, 2024; NPSA, n.d).

The paper examines Trusted Research through the theoretical lens of collaborative governance. In contrast to adversarial and managerial modes of public policy development and implementation, collaborative governance emphasises the role of public, private, and civil society stakeholders working together to achieve common policy goals or solve public issues (Ansell and Gash, 2008). Ansell and Gash define it as:

“A governing arrangement where one or more public agencies directly engage non-state stakeholders in a collective decision-making process that is formal, consensus-oriented, and deliberative and that aims to make or implement public policy or manage public programs or assets” (Ansell and Gash, 2008: 544).

This approach emphasises the importance of shared decision-making, where power and responsibilities are distributed among diverse actors who bring different expertise, perspectives, and resources, ensuring that policy solutions are co-created rather than imposed by a single authority. Collaborative governance often involves structured dialogue, transparency, and accountability to address complex issues that require cooperation beyond traditional government structures. It aims to foster trust among participants and enhance the legitimacy and effectiveness of policy outcomes (Ansell and Gash, 2007; Emerson et al, 2012).

The paper draws on semi-structured interviews with government officials, research funding organisations, university associations, senior university leaders, university research managers and individual scientists undertaken since August 2024 as part of an on-going study on international research collaboration and research security.

The paper will argue that the idea of collaborative governance can be applied to the Trusted Research programme. Since the launch of Trusted Research in 2019, the government has sought to foster partnerships between academic institutions, research funders and government agencies to safeguard sensitive research. University associations, especially Universities UK, have acted as an intermediary between government agencies and universities developing shared decision-making and communication channels among stakeholders to develop policies and procedures that balance the need for open scientific collaboration with national security concerns, such as intellectual property protection and the prevention of foreign interference.

Nonetheless, as the full paper will discuss, Trusted Research has been contested by the university sector, albeit behind closed doors. Universities are alert to the challenges to academic freedom. At the same time, behind the rhetoric of collaboration are very real fears about the reputational consequences of security breaches for universities and the danger of legal sanction for breaches of export control and other legislative provisions.

REFERENCES Adams J, Johnson J and Grant J (2022) “The rise of UK-China research collaboration: trends, opportunities and challenges”, Science and Public Policy 49:32-147. Ansell, C and Gash, A (2008) "Collaborative governance in theory and practice”, Journal of Public Administration Research and Theory, 18(4): 543-571. Emerson, K, Nabatchi, T and Balogh, S (2012) “An integrative framework for collaborative governance”, Journal of Public Administration Research and Theory, 22 (1): 1–29, https://doi.org/10.1093/jopart/mur011 G7 (2022) G7 Common Values and Principles on Research Security and Research Integrity, June 2022. National Protective Security Authority (2024) Trusted Research Guidance for Academics National Protective Security Authority and National Cyber Security Centre, available at https://www.npsa.gov.uk/system/files/trusted-research-guidance-for-academia-digital-july24.pdf National Protective Security Authority (n.d.) Trusted Research Guidance for Senior Leaders, National Protective Security Authority and National Cyber Security Centre, available at https://www.npsa.gov.uk/system/files/npsa-trusted-research-guidance-for-senior-leaders_0_1_0.pdf Quimbre, F, Carlyon, P, Dewaele, L and Drew, A (2022) Exploring Research Engagement with China: Opportunities and Challenges, RAND Europe: Cambridge. Universities UK (2022) Managing Risks in Internationalisation: Security Related Issues. Universities UK: London.

ABSTRACT. The "China Initiative," launched by the U.S. Department of Justice in November 2018 and concluded in February 2022, was a critical and contentious program intended to combat alleged national security threats posed by Chinese efforts to misappropriate U.S. intellectual property, especially within the realms of science and technology. This explorative study delves into cases centered around allegations of compromised research integrity, examining how a multitude of factors interacts to shape judicial outcomes and policy directions.The findings of this study reveal that no single factor is determinative of a guilty verdict; rather, outcomes are the result of complex interactions among various conditions. Key factors identified include the presence of financial ties to China, involvement in China's talent recruitment programs, the nature of the alleged misconduct, and the strength of the prosecution's evidence. The study stresses the need for a delicate balance between safeguarding national security and promoting international scientific collaboration.

ABSTRACT. Abstract submitted as part of the proposed thematic session: "As open as possible, as secure as necessary": A Discussion of the Securitization of World Scientific Research and Development

The global order is in flux amid intensifying competition for economic and technological advantage between the United States (US) and China. Concerns about the economic and security risks of dependency on China increasingly shape economic and political decision-making in the West and its closest allies in Asia. Against this backdrop, science and innovation are increasingly regarded as critical battlegrounds and sectors for national competitiveness. The trend is occurring within a global scientific landscape that has undergone dramatic shifts towards multipolarity since the 1990s, as an increasing number of countries partake in advanced scientific production, while the volume of publications has increased significantly. However, a multipolar world imposes substantial challenges to international exchanges.

Bibliometric data of collaborative patterns illustrates that international military alliances have not generally dictated science or technology cooperation. Data from 2022 shows that in the US the largest source of foreign research collaborators is China. Scopus data shows that for the European Union (EU), United Kingdom (UK), Australia, and Japan, collaborations with China were the second largest source of international research. These collaborative patterns indicate that factors such as scientific opportunities, resource complementarity, and individual relationships have been more important drivers of international scientific cooperation than geopolitics. However, national and global security concerns as well as nations’ global competitiveness are gaining precedence on the political agenda in many countries. China’s emergence as a global superpower challenging Euro-American hegemony and various liberal elements of international order has exacerbated global political tensions, particularly over science and technology. At the same time, the world is facing human-driven existential threats such as climate change, environmental deterioration, biodiversity loss, and pandemics that demand global collaboration and interaction between countries.

Existing literature has demonstrated how geopolitical tensions have negatively affected international scientific collaboration. Yet there has so far been few if any cross-national comparative examination of the patterns of variation across a broad set of countries. This paper investigates how countries respond to the changing geopolitical dynamics and manage the implications for their science sectors. Case studies include Australia, Japan, Sweden, the UK, and the US. These case studies illustrate a subset of countries that are aligned with respect to political affinity, security interests and level of economic development, but are different with regards to geographical location, size, level of openness and attention to research security. By investigating the variation in responses to international science collaboration, a deepened understanding can be gained of various practices used to promote international collaboration as well as the securitization of national research systems.

The paper will analyse the variation in responses taken at national levels building on the framework proposed by Shih et al. (2024). The framework suggests that within each national context, the nexus between open international research collaboration and competing national security imperatives varies along three key dimensions: actors (who responds); methods (by what means); and goals (to what ends).

Actors involved range from the local to global level. The response can also be left up to the individual. Alternatively, responses may be formulated collectively by university departments, universities central administrations, groups of universities, funders or sectors as a whole. Epistemic groups moreover play an important role such as peer networks, or academic societies in scientific boundary drawing. Finally, decisions involve the government as the main actor – from local authorities up to sanctions regimes, or discretionary prerogatives of national security agencies or the political executive.

Goals can range from ‘open science’ in which scientists operate completely freely from state influence to tackle common problems, to ‘national interest’, in which states exercise high levels of control over science to achieve goals such as national security, technological innovation, and economic development. In between lie the pursuit of numerous contested socio-political norms that are accepted to differing degrees in different contexts. Such norms include the utilitarian ethical principles, academic freedom and institutional autonomy, nationalism, individualism, and social justice.

Methods include the full array of diverse instruments and guidelines developed across national contexts. At one end of the spectrum is voluntary discretion, including the decisions of individual academics and collectives who voluntarily align with particular positions and practices. At the opposite end are regulatory frameworks that enforce compliance.

References

Shih, T., Chubb, A., & Cooney-O’Donoghue, D. (2024). Scientific collaboration amid geopolitical tensions: a comparison of Sweden and Australia. Higher Education, 87, pp. 1339-1356. https://doi.org/10.1007/s10734-023-01066-0

ABSTRACT. Introduction: PhD supervisors play a critical role in shaping the researcher identities of doctoral students, a process that is influenced by professional imprinting, a time-sensitive form of socialization (Marquis & Tilcsik, 2013). As doctoral students go through the early stages of their academic careers, supervisors use mechanisms to imprint specific norms, values, and expectations on their PhD students that ultimately become part of their researcher identity (Gruber et al., 2024; Heron et al., 2023). However, this development process is complicated and influenced by many factors, among others, gender and diversity, which can either facilitate or inhibit the growth of a resilient and heterogeneous researcher identity. In this line, women and scholars from underrepresented backgrounds, for example, often face challenges within academia, like the lack of role models, unconscious biases, gender pay gaps, family responsibilities etc. (Beaudry & Larivière, 2016; Eslen‐Ziya & Yildirim, 2022; Howe-Walsh & Turnbull, 2016; Ivanova-Gongne et al., 2021; Järvinen & Mik-Meyer, 2024). These challenges may result in specific pronunciations of researcher identities that remain underinvestigated until today. Our study addresses the gap in understanding how PhD supervisors influence researcher identity development through professional imprinting with a focus on gender and diversity, putting forward the following research question: How do PhD supervisors influence the development of researcher identities through professional imprinting, focusing on gender and diversity?

Theoretical Background: The development of researcher identities is informed by social identity and identity theory. These closely related perspectives emphasize how individuals construct identities through social interactions (Hogg et al., 1995). Identity theory focuses on the self-meanings that emerge from distinct roles, while social identity theory highlights belonging to groups within various contexts (Stets et al., 2020). In academia, a PhD student’s researcher identity evolves through adapting to institutional environments, giving meaning to certain roles and seeking validation from peers and mentors through interaction with social networks, a process also known as ‘socialization’ (Coombs, 1978; Gruber et al., 2023; Ivanova-Gongne et al., 2021; Pretorius & Macaulay, 2021; Seyri & Rezaee, 2022; Tierney, 1997). Professional imprinting, a time-sensitive, intense early-stage socialization, occurs in periods of increased susceptibility, i.e., at the beginning of a career, when individuals are especially receptive to influence (Marquis & Tilcsik, 2013). Supervisors play a significant role in this process, imprinting values, expectations, and norms on PhD students, which become integral to their researcher identities. Scholars have shown that supervisors’ entrepreneurial engagement can influence individuals’ entrepreneurial behaviors, illustrating imprinting at work (Azoulay et al., 2017; Bercovitz & Feldman, 2008). The PhD student-supervisor relationship is crucial in this context, as it directly impacts the success of the doctoral journey (Bastalich, 2017; Johansson & Yerrabati, 2017; Sverdlik et al., 2018). This relationship shapes students' research orientations, e.g., favoring theory-focused research and/or commercially oriented research, creating a spectrum of researcher identities (Gruber et al., 2024; Roach & Sauermann, 2010; Sauermann & Roach, 2012). Finally, Gender and diversity significantly impact researcher identity development, shaping access to mentorship, socialization, and identity validation. Women and underrepresented groups often face systemic barriers in academia, such as limited mentorship and implicit biases, which can hinder their ability to form a strong researcher identity (Griffin et al., 2020; Wofford & Blaney, 2021). Scholars from diverse backgrounds can bring unique perspectives or research topics, but these can sometimes clash with existing academic norms, leading to challenges in achieving a sense of belonging (Heron et al., 2023; Johnson, 2019). Additionally, research shows that scholars from underrepresented groups often navigate dual identities, balancing personal and academic expectations (Choi et al., 2023; Holley, 2015), leading to additional barriers in the trajectory.

Data and Method: We launched a survey distributed to PhD students and postdoctoral researchers across Europe and Canada. We use a non-probability purposive sampling method. So far, we have received 485 completed responses, but the data collection is still ongoing. The survey gathers data on participants’ demographics as well as their perceptions of what constitutes a “good researcher”, mobility and career aspirations. Respondents are asked to reflect on their views before starting their PhD journey and during or after and to assess the importance of 14 distinct aspects. Some of these aspects are associated with more ‘traditional’ research practices, such as publishing in high-impact journals, presenting at conferences, building a scientific reputation, and working autonomously, others are associated with more ‘commercial’ research practices, such as patenting, industry-collaborations or impact for society. We then asked as well how much the participants believed their supervisors influenced these practices. We also explore whether other individuals, such as peers, family, or industry partners, played a role in shaping these views. Initial descriptive analysis and data exploration using a two-step clustering technique will be performed. The latter technique categorises individuals into coherent clusters that show similar patterns in key features of their responses. After validating the different clusters, we will use conventional statistical methods, like regression and correlations to analyse those patterns. Gender, diversity, field of study, country of study, etc. are all taken into account at that stage. Ultimately, further analysis will characterise the different researcher identity development trajectories and distinguish the influence of the supervisor and other individuals on this development.

Results and Implications: With this study, we present the need for policy reforms in academia to support the development of heterogeneous PhD student researcher identity, considering the importance of gender and diversity. Our findings will also suggest that universities should adopt more flexible definitions of academic success, valuing both traditional and nontraditional research pathways, e.g., moving to industry. This shift would benefit diverse scholars who often bring unique perspectives to academia. By highlighting the influence of supervisors on identity formation, this research equally points to the value of inclusive mentorship training that encourages supervisors to address the unique challenges faced by underrepresented groups. Additionally, increasing career development resources that support different career aspirations and bring consciousness to identity development can help PhD students balance personal and professional growth and mitigate identity struggles that often lead to the PhD abandonment. Policy adjustments in academia can create a more inclusive culture that values diverse research identities.

ABSTRACT. Critical to academic training is the relationship between students and supervisors. In this relationship, the learning process has been traditionally viewed as the supervisor unidirectionally transferring knowledge to the student (Delamont and Atkinson, 2001). However, unidirectionality is not always the case. Students may acquire specific knowledge faster than their supervisors (Stephan and Levin, 1992) and transfer that knowledge back to them. This is especially true in fast-changing technological landscapes (Fleming and Sorenson, 2003), in which supervisors often learn from their students the use of new technologies. One such technology is Artificial Intelligence (AI). The use of AI in science has grown fast in the last decade, transforming scientific research across research fields (Flanagan et al., 2023). Given its recent upsurge, students might have better and more up-to-date AI knowledge than their supervisors, becoming a source of AI knowledge. The goal of this study is to provide a comprehensive understanding of academic training by illustrating the bidirectionality of knowledge transfer between supervisors and students.

To empirically document the AI knowledge transfer process, we draw on a unique dataset covering the entire population of 51,826 French PhD students in STEM who graduated between 2010 and 2018. We focus on AI as the area of knowledge transferred for two main reasons. First, AI is a contemporary and growing technology (Flanagan et al., 2023), and advanced economies, including France, have made significant investments in research and education concerning the technology. Second, AI is a general-purpose technology (Cockburn et al., 2018) applicable in all fields of science, and thus, the learning of the technology has increasingly become critical.

In our empirical analyses, we test the existence of a supervisor-to-student knowledge transfer and the student-to-supervisor knowledge transfer using the following approach.

Supervisor-to-student knowledge transfer To analyze the supervisor-to-student knowledge transfer, we identify the student-supervisor pairs in which the supervisor has AI knowledge before the student’s enrollment (〖AI supervisor before〗_i=1). To address selection mechanisms that might intervene and bias our estimates, for each pair having a supervisor with AI knowledge, we find a similar pair having a supervisor with no AI knowledge before the student’s enrollment (〖AI supervisor before〗_i=0). To do so, we employ a Propensity Score Matching (PSM) approach based on the student-supervisor pairs’ observable characteristics. We compare the probability of writing an AI thesis (〖AI student〗_i=1) for students in student-supervisor pairs having an AI supervisor to the probability in a control sample of similar pairs having a non-AI supervisor. Our findings reveal a statistically significant difference in outcomes based on the supervisor's AI background. Specifically, pairs in which students are mentored by AI supervisors exhibit a 16% probability of showing a student writing an AI-related thesis, compared to the 4% probability in pairs in which non-AI supervisors mentor students. The statistically significant difference between the two probabilities, i.e., 12 percentage points, indicates an AI knowledge transfer from the supervisor to the student.

Student-to-supervisor knowledge transfer To analyze the student-to-supervisor knowledge transfer, we isolate the student-supervisor pairs in which supervisors have no previous AI knowledge before mentoring the students. We identify the student-supervisor pairs in which students develop AI knowledge during their thesis (〖AI student〗_i=1). Then, for each of these pairs, we find a similar pair characterized by a student with no knowledge in AI (〖AI student〗_i=0) using a PSM approach. Finally, we compare the probability of a supervisor initiating AI research within the three years after their student's defense year (〖AI supervisor after〗_i=1) for supervisors in student-supervisor pairs having an AI student to the probability in a control sample of similar pairs with a non-AI student. Our results reveal a statistically significant difference in outcomes. For pairs in which the student writes an AI thesis, the probability of the supervisor publishing in AI within the three-year window following the student's defense is 33%, compared to a lower probability of 14% for supervisors in pairs where the student does not write an AI thesis. The statistically significant difference between the two probabilities, i.e., 19 percentage points, indicates knowledge transfer from the student to the supervisor.

Our results are relevant for policymakers who want to steer the research agenda of a country toward specific topics. They show two main channels through which AI can permeate the research system of a country. One channel is to incentivize established researchers to develop AI knowledge, which they can transfer to their students. Another channel is the AI education of young generations of undergraduate students. Indeed, our results show that individuals acquiring AI knowledge before enrolling in a PhD program or during the PhD period can transmit this knowledge to their supervisors, steering supervisors’ research efforts toward AI.

ABSTRACT. Aims and Research Questions： This study empirically examined whether fellowship programs are effective in fostering talented young researchers. Specifically, it focuses on the Japan Society for the Promotion of Science (JSPS) Postdoctoral Fellowship (JSPS-PD), a typical program that supports young researchers before and after obtaining their PhDs. The main objective is to clarify the impact of receiving a fellowship on the research productivity of researchers and, in particular, to examine this from the perspective of the number of academic papers published and the number of citations received. Furthermore, this research also explores how factors such as gender, researcher mobility, speed of finding employment after being awarded a fellowship, and overseas experience affect research productivity. The specific research questions are as follows: (1) Does receiving a fellowship improve research productivity? (2) Is research productivity higher for those who find employment quickly after being awarded fellowships? (3) Does overseas experience improve research productivity? (4) Does research productivity increase for researchers who frequently change their affiliations? This study focuses on researchers in the humanities and social sciences, including economics, where the effects of fellowship programs have not been studied as extensively as in the natural sciences. However, the core mechanisms of fellowship support - financial security, resources, and time for intensive research - are relevant across various disciplines. Therefore, the findings and methods of this study are expected to provide universal insights that will help policymakers design effective fellowship programmes.

Method： A comprehensive microdataset was created to address these research questions, including data from 236 fellowship recipients and 454 non-recipients. The propensity score matching method was used to pair researchers with similar educational backgrounds and personal attributes, minimizing bias and enabling fair comparisons. The main variables analyzed included personal characteristics (e.g., gender, number of transfers between institutions, time to first job after fellowship adoption, overseas research experience, type of university/graduate school, etc.) and research performance (e.g., number of English papers published, number of citations, number of Japanese papers published, etc.). Academic paper data was obtained from databases such as “Web of Science” and “CiNii”, and information on doctoral degrees was obtained from databases such as “CiNii Thesis Information” and “ProQuest”, while details of career history were collected from publicly available CVs. A regression model was used for the analysis, and the impact of fellowships on subsequent research productivity was evaluated while controlling for potential confounding factors, such as past research achievements, academic background, and gender. This approach was essential for isolating the effects of fellowships and accurately evaluating their impact on the research results without distortion by external variables.

Results： The results clearly show that receiving a fellowship increases research productivity, particularly in terms of English publications and citations. On the other hand, the impact of fellowships on publications in Japanese is limited. Therefore, it is suggested that the JSPS-PD program is effective in fostering young researchers who can play an active role internationally. Furthermore, we found that researchers’ mobility is positively correlated with improvements in research productivity. Researchers who move between institutions tend to have a wider network and access a greater variety of resources, which is likely to contribute to improvements in research results. We also found that fellows who could find employment quickly after receiving their fellowship were more productive. This indicates the importance of stability during the early stages of a researcher's career. Interestingly, while it is generally thought that research experience abroad is beneficial, this analysis did not find a consistent positive correlation with increased productivity. The results of this survey call into question the common belief that international experience improves research performance and suggest that other factors, such as the quality of the research environment, may be more important. The survey also reveals a significant gender gap. Female researchers tend to publish fewer English-language papers than their male counterparts, possibly because career interruptions related to personal life events such as marriage and childbirth often coincide with the fellowship period. Despite the support provided by fellowships, female researchers may face unique challenges that hinder productivity. To address this issue, more targeted support mechanisms are needed to mitigate the effects of career interruptions and enable female researchers to pursue their academic goals without being disadvantaged.

Conclusion： In conclusion, this study provides robust evidence that fellowship programs such as the JSPS-PD effectively enhance the productivity of young researchers, particularly in an international context. Not only do recipients publish more papers but they also achieve higher citation counts, indicating that the program is successful in developing researchers from a global perspective. The importance of early career stability was also revealed, with the results showing that fellows who found employment quickly after being awarded a fellowship were more productive. However, the results showing that there were no consistent benefits from overseas experience suggest that simply encouraging people to gain international experience is not sufficient. Furthermore, the observed gender gap highlights the need for tailored support systems to address the unique challenges faced by female researchers. Although this study focuses on economics researchers, the insights are broadly applicable to a variety of fields. The main benefits of fellowships - financial security and the time to focus on research - are essential for young researchers across disciplines. These findings provide valuable guidance for policymakers in maintaining fellowship programs as an important tool for cultivating the next generation of academic talent and building effective support systems for early career researchers.

ABSTRACT. Research training programs are essential for developing the next generation of researchers, because they provide the knowledge, skills, and contacts necessary for young scientists to become successful independent investigators. Due to their importance, biomedical research training programs supported by the US National Institutes of Health (NIH) are frequently evaluated, but such evaluations are typically limited in scope for at least two reasons. First, they typically focus on a single program or training type and include fewer than 100 trainees. These evaluations can provide useful information for the specific program being evaluated, but their findings are difficult to generalize or compare with other programs. Second, most of these evaluations focus on subsequent NIH funding as the primary measure of trainee success. Although subsequent NIH funding can be an important step in becoming an independent biomedical investigator in the US, reliance on this measure undervalues other successful trainee outcomes.

In this presentation, we introduce a new method of obtaining trainee outcome data that addresses both limitations. We introduce a fully automated method of linking trainees to their Scopus author profiles, allowing us to obtain information on the current affiliations and research activities of thousands of NIH trainees from all NIH-supported training types. This also allows us to introduce a new measure of trainee success: whether the trainee is still actively publishing in the academic literature, and therefore still engaged in a research-related career. We then run this procedure on all trainees participating in all training programs supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) from fiscal year 2000 through 2019. We then compare the traditional funding measure to the new active publishing measure for these trainees overall and by training type.

The proposed method has five steps. First, we obtain training award data and data on subsequent NIH applications and awards from our internal NIH data systems and perform some standardization and processing steps to ensure accurate matching. Second, we search the Scopus Author APIs for each trainee by name and affiliation at the time of training and retrieve up to three possible matches. Third, for trainees not matched in step 2, we search the APIs by name and subject area and retrieve up to five possible matches. Fourth, we use the APIs to obtain all Scopus Author profiles for all possible matches. Finally, we run a custom decision algorithm to select the most probable author profile for each trainee. Accuracy testing of the method against manual searches for 200 randomly selected trainees indicated that the method achieves 91% recall and 97% precision.

A total of 10,613 individuals participated in NICHD training programs in FY2000-2019. Approximately 20% of these trainees received one or more subsequent research project grants from the NIH, with that proportion remaining relatively constant across the years in our analysis. However, these rates varied substantially by training type, with our graduate and immediate postdoctoral training programs having lower funding rates (at around 15% and 30%) than our career development programs (at around 40% and 60%). Training programs awarded to institutions had lower rates than programs awarded directly to individuals across all trainee career stages. Individuals receiving student loan repayment awards had subsequent funding rates (at around 50%) comparable to those of individuals participating in career development awards, even though loan repayment awards do not include any formal training or mentoring.

Approximately 89% of NICHD trainees were successfully matched to a Scopus author profile. Match rates were higher for career development trainees than for graduate and postdoctoral trainees. Approximately 52% of all trainees in our analysis are actively publishing, defined here as authoring or co-authoring at least one peer-reviewed publication in 2023 or 2024. As with subsequent funding, active publication rates varied substantially by training type, with graduate and postdoctoral programs having rates of approximately 44% and 56%, respectively, and career development programs having rates around 80%. Loan repayment recipients had a rate around 72%. Approximately half of the trainees in our analysis who are actively publishing have not applied for subsequent research project funding from the NIH.

These findings have important implications for NIH training programs. First, we demonstrate that the proposed method can be successfully scaled to obtain outcome information for large numbers of trainees, making it a useful tool for subsequent analyses of other training programs at the NIH and beyond. Second, we find a substantial difference between the two outcome measures across all years and all training types in our analysis. Focusing on subsequent NIH funding as the primary measure of success therefore substantially underestimates the success of NIH training programs. Broader outcome measures are needed to adequately evaluate these programs. Third, we find that trainee success rates vary by career stage, with early career trainees having lower success rates than later career trainees. This suggests that factors external to the training are affecting these rates and that training program success should therefore be defined differently at each trainee career stage. Finally, these results suggest that approximately one-third of NICHD trainees are actively publishing but have not directly received any NIH research project funding. More research is needed to understand why these trainees have not received NIH funding and what might be done to support them in doing so.

ABSTRACT. Introduction Publicly funded researchers are increasingly under mandates to make their research outputs “open” (Dalipi et al., 2022, and others). However, some outputs arising from publicly funded research can be commercialised, which argues for the need for closure when intellectual property/patents are at stake (O’Mahony & Bechky, 2008). Indeed, some of the funding programs that require open research are also designed to foster commercial activity. The tensions between open research and intellectual property are particularly complicated when research is carried out by public-private research partnerships [PPRP] - i.e., projects between universities and other private sector partners (Perkmann & Schildt, 2015; Lattu & Cai, 2022).

As Cai and Lattu (2023) and others note, even as openness in science has long been an idealized norm of scientific research, such norms have been increasingly been complicated both by the changing nature of research, the values of different research actors/stakeholders, and the pervasive integration of digitalization in all aspects of the research process. In this paper, we explore how tensions among norms of openness and closure are understood and navigated by research ecosystem actors – researchers themselves, technology transfer offices, research support units, funders, and others – in the Irish higher education and research context. In particular, we also add to the growing body of work on this topic by integrating the perspectives of data stewards and data managers, whose role has been ignored or eliminated in the context of commercialization. We also draw on policy and practice documents from funders, higher education institutions, and other national and supranational bodies (such as the Organization of Economic Cooperation and Development, or the OECD) to provide the context in which open science and commercialisation of research are done.

Analytical Framework We draw heavily on the work of Lattu and Cai (2022) who conducted interviews with research stakeholders in Finland and China to explore the tensions of university-industry partnerships. They frame their work in the context of Institutional logics that have been used in organizational studies to explore how individuals navigate complex institutional relations – and PPRP are one example of such a complex institutional system. Institutional logics encompass the norms, values, belief systems, histories, and material practices that underpin and constitute actions. In their work, institutional actors are generally found in their work in academic, industry, and government and intermediaries whose role is to facilitate collaborations across these sectors. The analysis of competing institutional logics forms the heart of their analyses.

Thornton and Ocasio (1999, pg. 804) provided a definition of institutional logics as ‘the socially constructed, historical patterns of material practices, assumptions, values, beliefs, and rules by which individuals produce and reproduce their material subsistence, organize time and space, and provide meaning to their social reality’. Logics are believed to provide guidelines on how to interpret social situations and compiling with them provides legitimacy and confirmation for organisations by their society (Thornton & Ocasio, 2008).

In academic research, conflict between the logic of open science and commerce may arise when universities pursue open access and where this conflicts with the pursuit of commercialization of research results. Researchers of institutional logics seek to understand the organizational responses to these tensions and this is at the heart of our paper. Furthermore, some of the professionals with the greatest interest and expertise in open science – the data stewards and data managers who work in higher education, funders, and other bodies - are not considered as part of the PPRP ecosystem in the literature. In this paper, we look at the competing institutional logics and the tensions inherent in them and also look at how data stewards and managers are conceptualized by these other actors and how their institutional logics play out.

Research Design We draw on Irish and international policy documents from funding bodies, universities, and other organisations with relevant interests. Themes which were pursued in semi-structured interviews with researchers and research support staff. We did not pursue a systematic approach to recruitment but instead drew on convenience sampling for this study. We also conducted a survey of data stewards and others with similar roles. Because there is no standardization with respect to the titles of these roles, we disseminated our survey via a listserv of individuals working in data stewardship across Ireland. We received 19 responses which comprised both closed- and open-ended questions. Data were analysed based on Clarke and Braun’s concept of Thematic Analysis (2017) using Atlas.ti.

Findings Interviewees and survey respondents cited lack of time, insufficient support for making decisions around when/if to make work open (or commercialise it), and potentially sharp distinctions between academic and commercial work, and lack of communication among stakeholders. Some data stewards cited downright hostility between themselves and the technology transfer offices at their institutions; researchers felt that they were not given clear pathways and guidance, especially with respect to commercialisation. The research team will expand upon the number and type of interviews and also delve further into policy analysis to learn more about open research and PPRP.

Bibliography

Clarke, V., & Braun, V. (2017). Thematic analysis. The journal of positive psychology, 12(3), 297-298.

Dalipi, F., Ferati, M., Kurti, A., & Kastrati, Z. (2022, June). Investigating the FAIRness of Science and Technology Open Data: A Focus in the Scandinavian Countries. In International Conference on Human-Computer Interaction (pp. 276-283). Cham: Springer International Publishing.

Lattu, A., & Cai, Y. (2023). Institutional logics in the open science practices of university–industry research collaboration. Science and Public Policy, 50(5), 905-916.

O'Mahony, S., & Bechky, B. A. (2008). Boundary organizations: Enabling collaboration among unexpected allies. Administrative science quarterly, 53(3), 422-459.

Perkmann, M., & Schildt, H. (2015). Open data partnerships between firms and universities: The role of boundary organizations. Research Policy, 44(5), 1133-1143.

Thornton, P. H., & Ocasio, W. (1999). Institutional logics and the historical contingency of power in organizations: Executive succession in the higher education publishing industry, 1958–1990. American journal of Sociology, 105(3), 801-843.

Thornton, P. H., & Ocasio, W. (2008). Institutional logics. The Sage handbook of organizational institutionalism, 840(2008), 99-128.

ABSTRACT. Academic patents are one of the most important means of technology transfer from science to industry. Technology transfer is in the focus of innovation research since many years. The reason is that technology transfer can be seen as means to increase the technological competitiveness of companies and of nations, but also as means to pay back the investment made by the so-called taxpayer into science. In other words, next to private returns to the commercializing company, public value is also expected to emerge from technology transfer. The work presented here offers a reliable method to identify academic patents, i.e. patents invented by a researcher affiliated to a university plus patents filed by the university itself - at large scale for the US and Canada. A similar method was successfully implemented for 39 EPC member countries. The analyses of the so generated data show considerable differences in the ownership patterns of academic patents across countries, mainly reflecting the existing/missing regulation of ownership (e.g. Bayh-Dole). This paper will extend the existing method to the US. While it has been successfully implemented for Canada already, the implementation of the method for the US is pending. The core challenge for the US is the larger science community, leading to larger numbers of homonyms. In addition, for US researchers/inventors the distance between work and home addresses might be larger than in other countries, posing a challenge for the method that uses geographic distance as one means to disentangle the homonym problem. Next to the implementation of the method itself, validation needs to be undertaken as well. The first research question is, if it possible to extend the existing methodology to the US at a reasonable quality (precision and recall) of the data. Second, if the methodology can be implemented successfully, we will analyse the trends of academic patenting in North America and compare the structures to selected countries in Europe. One way of technology transfer might be based on licensing contracts. However, licensing contracts are private and usually not disclosed and therefore hard to identify or quantify. Other means or indications of technology transfer are, for example, contract research, joint publications, informal knowledge transfer, or spin-offs. Academic patents might be an outcome of joint research, of successful licensing or of ownership transfer. Academic patents are defined as patents filed by the university and/or invented by at least one person affiliated with the university and filed by another entity. The USA was the first country to introduce a regulation - called Bayh-Dole act - that allows universities the (first) ownership and decision-making on inventions emerging from government-funded research. Identical or similar regulations have been introduced in several countries afterwards. For example, the professor's privilege was abolished in Germany in 2002 till then allowed a professor to file and own the patent rights emerging from inventions made by her. Since 2002, the university owns the invention and decides (first) on its use. Several other countries, for example France, Sweden or Italy (see Lissoni et al. 2009), have also implemented Bayh-Dole-Act-like regulations in the early 2000s, while Canada still has no such regulation implemented. The numbers of university-filed (owned) patents increased in most countries that have introduced such a regulation. The methods underlying the data identification activities in this paper aim at the identification of the (complete) patent output of universities (academic patents). The challenge, however, is to identify university-invented patents as one part of academic patents. To do this, a matching procedure of author names from Elsevier's Scopus with inventor names from PATSTAT will be performed for the US. Next to missing inventors in Scopus (not all academics who file a patent might also publish papers in journals that are listed in Scopus), the biggest challenge of this procedure is the treatment of homonyms. Our algorithm follows a rule-based approach that builds on a Levenshtein distance string matching algorithm of author/inventor names, with a matching score of 0.89 that was identified as an optimal level in a gold-standard dataset of 200 inventor-author linkages. In addition, we allow matches only if the patent and the publication are within a similar time window, i.e. the priority date of the patent should be in the same year as the publication or the previous year. Furthermore, we accept an author-inventor match only if the addresses are within a certain distance. For this purpose, we geocode both, patents and publications. As distances between home and work addresses of researchers in the US might be larger than in Europe, we need to develop an optimal distance that fits the US context. Hence, we will create another gold-standard dataset for the US. To further reduce homonyms (false positives), the patent and the publication are required to fall into similar technological areas, which we assign by a coarse-grained concordance of technology fields in patents and scientific fields in publications. The outcome of this procedure is a dataset of patents flagged as academic patents. Once the dataset is established, we intend to analyse the trends of academic patenting in North America and compare them with the trends that we found in our study on Europe. To be able to analyse the applicant structure of university-invented patents, we resort to our PATSTAT-ORBIS link.

ABSTRACT. Situated in the dual-disadvantaged position of being in a resource-constrained institution in a developing country, we conducted a pilot-study to examine how academics’ consultancies with commercial entities impact problems in research, mentoring practices, and research productivity. Our findings derive from analyzing original face-to-face survey data using bootstrap regression and principal component analyses. Three core problems in research emerged: management-, instrument-, and coordination-related. Instrument-related problems were mitigated by involvement in consultancies. Regarding mentoring practices, three emerged: socializing students with commercial entities, providing students with guidance and feedback, and helping students with job search. Consultancies were directly linked to the socialization of students to commercial entities, which included exposing them to consultancies, collaborations, and funding opportunities. Involvement in consultancies and volume of productivity were not directly linked. The socialization of students to commercial entities linked consultancies to productivity. Our results provide an insightful view of and generate hypotheses on how academics’ consultancies shape mentoring and problems in research. Contrary to claims that academics’ engagements with commercial entities are ‘malignant’ to science -- especially so in the context of resource-constrained institutions -- our pilot results indicate that such engagements do not translate to dismal outcomes, exacerbate research problems, or generate irregular mentoring practices.

ABSTRACT. The aim of this presentation is to explore the historical emergence of the science of science, specifically drawing from its development in Poland, and to demonstrate its relevance for shaping contemporary science policy. The talk will shed light on how historical lessons from the formative years of the science of science can inform current debates and approaches to science governance. This historical analysis serves not only as a testament to the contributions from peripheral scientific communities but also as a guide for modern frameworks seeking to balance practical and theoretical insights in science policy.

The emergence of science of science

Rapperswil Castle from the early 13th century located on a peninsula on Lake Zurich witnessed centuries of history before falling into disrepair. In 1870, it found a new lease on life when Polish émigré Władysław Ewaryst Plater leased it and transformed it into the Polish National Museum. Half a century later, plans were underway for Rapperswil Castle to become the home of the world’s first Institute of Science of Science, to be led by Maria and Stanisław Ossowski. Though this ambitious vision never materialized, it underscored the superior aspirations of Stanisław Michalski, the driving force behind this science-of-science endeavor.

Today, Michalski’s name might not resonate with many historians of science, even in Poland, let alone beyond its borders. Born in 1865, Michalski was a vital figure in the scientific movement that, in the early 20th century in Warsaw, led to the emergence of a new academic discipline: the science of science. He also served as the founding editor of the world’s first strictly science-of-science journal, Nauka Polska. Jej Potrzeby, Organizacja i Rozwój (en. Science and Letters in Poland: Their Needs, Organization, and Progress). This journal in 1935 published a cornerstone work for the new discipline, namely an article by Maria and Stanisław Ossowski. Their work laid out the first comprehensive program for this new discipline designed to enrich both science and society with its theoretical and practical insights.

Why did the science of science emerge specifically in Poland?

The shortest possible answer is this: a group of Polish scholars, due to objective circumstances, was primarily educated outside the borders of what was then a non-existent Poland on the map of Europe. They participated in international research and discussions on the status and role of science. In terms of understanding what science is, how it should be practiced, and its role, there was nothing particularly unique to explain the emergence and development of the science of science in Poland. What was unique, however, at the turn of the 19th and 20th centuries, was the end of Poland’s partition into three parts in 1918 after 123 years of non-independence. This challenge of unifying the three partitions and understanding the potential role of science in this task created the historical conditions for proposing the science of science in Poland. Scientists in no other country, who also discussed science and its role, faced the same political, cultural, and societal task as Polish scientists in similar historical circumstances. However, it should be noted that science was perceived as a significant hope for technological and political development in many of the new states founded in the former imperial spaces of Central and Eastern Europe after the First World War.

Reintegrating these three separate regions, each with different administrative, legal, and educational systems, infrastructure, health system, and units of measurement into one cohesive state was an unprecedented challenge. These experiences equipped Polish scholars with a broad and comparative perspective on the role of science in society. They understood that science was not just a pursuit of knowledge but a crucial instrument for national integration and modernization. This perspective was essential to the emergence of the science of science in Poland, emphasizing the need for a reflective and organized approach to scientific practice and policy. Thus, science of science as metadiscipline was born out of the necessity to integrate disparate scientific traditions and leverage science as a tool for national development and identity reconstruction.

Science as a policy tool for national integration and modernization

Since the publication of the Ossowskis’ work, much has been written about science studies and the science of science itself. When re-reading their “Science of Science,” it is crucial to remember that their approach—treating science as a social and cultural phenomenon—was far from obvious at the time. In fact, it was quite revolutionary during the 1920s and 1930s.

The Ossowskis emphasized viewing science as a collective human endeavor, moving away from the traditional focus on isolated individual achievements. They were pioneers in advocating for an interdisciplinary approach to science studies, recognizing that a comprehensive understanding of science necessitates integrating insights from philosophy, sociology, psychology, and the practical organization of scientific activity. This integration, they argued, should be collaboratively undertaken by the scientific community. Today, their interdisciplinary vision is increasingly relevant, as modern science studies and scientometrics employ diverse methodologies to tackle complex questions about the nature and function of scientific work. A particularly forward-thinking aspect of the Ossowskis’ program was their focus on the practical and organizational dimensions of science. They argued that effective scientific practice and policy require systematic study and understanding, analogous to the planning and analysis required in industrial production. In today’s context, where research evaluation, science policy, and the management of scientific institutions are central concerns, their emphasis on practical applications remains highly applicable. The Ossowskis anticipated many of the challenges that modern science governance faces today, making their insights invaluable for contemporary discussions on how to manage and support scientific research effectively.

Moreover, the Ossowskis emphasized that the way we conceptualize science directly influences the training of future generations of scientists and how they perceive the role of science in society. Revisiting their program can inspire new ways of thinking about how we study and evaluate science, serving as a reminder that our current approaches are built on a rich history of reflection and debate.

ABSTRACT. Introduction

The “science of science” (SciSci) has established itself as an emerging interdisciplinary academic field; related research uses big data and computational technologies to examine the processes and mechanisms by which scientific knowledge is newly created, shared, and later institutionalized through a complex, self-organizing, and evolving network of scholars, projects, papers, and ideas (Fortunato et al., 2018). In tandem with established academic fields such as the history of science, philosophy of science, sociology of science, scientometrics, library and information science, and science for policy, SciSci has deepened our understanding of how science succeeds by quantitatively examining the process by which various scientific agents interact across diverse geographic and temporal scales (Wang & Barabási, 2021).

SciSci’s progress has offered us evidence-based findings that help promote science, technology, and innovation policy. However, this only applies to the science of “hard science”; research on the science of “soft science” (i.e., the humanities and social sciences) is far less extensive. Thus, we know little about the processes and underlying mechanisms by which scholars in the humanities and social sciences share their ideas, collaborate, write papers and books, and engage in social activities.

To fill this research gap, we first need to capture how scholarly and social activities in the humanities and social sciences may have a social impact. Shedding more light on the science of soft science and its impact on society could increase our understanding of how the creation and adoption of new ideas, knowledge, and technologies evolve over time alongside individual values and psychologies, societal values, business models, political relations, legal and policy frameworks, and regulatory regimes.

This paper aims to empirically examine this social impact in terms of direct and indirect effects on science, technology, and innovation policy. The first task is to identify different types of scholars according to their professional identities among disciplines by conducting qualitative interviews with scholars about their professional identities. The second task is to examine differences among the various humanities and social sciences by quantifying scholarly and social activities and their academic and non-academic outputs.

Sample and Methods

To empirically examine the characteristics of scholarly and social activities among humanities and social sciences scholars, we first conducted a qualitative study (Study 1) and then conducted a quantitative study (Study 2). In Study 1, we conducted semi-structured interviews with 78 scholars randomly selected from 41 of the 69 humanities and social sciences subfields defined by the Ministry of Culture, Sports and Education in Japan. The interviews were conducted from December 2023 to July 2024. Interviews lasted on average from 60 to 90 minutes; in total there were 84 hours and 19 minutes of interviews. Every interview was audio-recorded, transcribed, and coded to identify issues related to the participants’ research activities, environment, research outcomes, research evaluations, and subjective relationship to society.

After coding qualitative data for Study 1 based on the grounded theory approach, Study 2 was conducted to identify similarities and differences among scholars by types of organization, main fields, and sub-fields. Study 2 was based on a bibliometric examination which quantified the numbers and frequencies of papers, books, academic presentations, reports, and various kinds of academic and non-academic activities for 95753 scholars. We utilized a national-level Internet database service called Researchmap, run by the Japan Science and Technology Agency (Arai & Masukawa, 2010). Researchmap is an Internet database service that collects information on researchers, including their career histories and lists of papers; it includes data relating to Japanese researchers and foreign researchers affiliated with Japanese research institutions (including national, public, and private universities; graduate universities; junior colleges; technical colleges; and research institutions). We merged individual activities and their outcome by year, types of affiliated organizations, and 69 subfields consisting of 11 main fields (psychology, political science, geography and anthropology, economics and business administration, social and economic agriculture, education, literature and linguistics, history and archaeology, philosophy and art, sociology, and law).

This research represents the first attempt to comprehensively quantify research activities and outcomes of scholars in humanities and social sciences in Japan. These data, encompassing peer-reviewed articles and broader research activities, including participation in government councils, engagement in mass and social media, and involvement in archaeological excavations, offer key insights.

Empirical Findings

The empirical findings of this paper are summarized in the following two points. First, we identified similarities and differences among humanities and social sciences disciplines regarding research activities, research output, and social engagements. International activities and publications in international refereed journals were more common among psychology and economics scholars. In contrast, publishing in non-refereed publications and Japanese-language books, as well as conducting local activities, were more common among literature, sociology, and law scholars.

Second, the degree of social engagement differed among scholars, though not according to their disciplines; this suggests differences among scholars in terms of their socio-cognitive definitions of their professions. That is, such differences appear to depend not only on the universal nature of academic professions but also on historically determined institutionalized divisions of labor. Although our findings have some limitations, they can contribute to research examining the notion of a National Innovation System.

References Arai, N., Masukawa, R. (2010) Researchmap opening the door to the world of Science 2.0, Proceedings of the 13th IASTED International Conference on Computers and Advanced Technology in Education, CATE 2010, pp.166-171. Fortunato S., Bergstrom, C.T., Börner K., Evans, J.A., Helbing, D., Milojević, S., Petersen, A.M., Radicchi, F., Sinatra, R., Uzzi, B., Vespignani, A., Waltman, L., Wang, D., Barabási, A.L. (2018) Science of science, Science, 359 (6379). Wang D, Barabási A-L. (2021) The Science of Science. Cambridge University Press.

ABSTRACT. This study investigates whether value expressions (VEs) in patent documents - narrative statements describing societal or commercial problems, solutions, and promises - reflect the technological orientation (i.e., the potential to address certain commercial or societal problems based on the invention's objective technological features), or primarily serve as rhetorical devices. The research focuses on patents in artificial intelligence and nanotechnology, two fields with significant potential for societal impact (Roco, 2023; OECD, 2024; Goos and Savona, 2024). Using both generative and discriminative large language models, we analyze 175,730 U.S. patents filed between 2005 and 2023, categorizing statements into three types: public value expressions (PVEs) related to societal impact, private value expressions (PRIVEs) related to commercial value, and combined expressions (PVE-PRIVEs) covering both aspects. Specifically, we implement a method based on Pelaez et al. (2024), which involves using a structural prompt with GPT-4 to label 20,000 samples. These labeled samples then serve as a training set for fine-tuning RoBERTa Large, enabling it to predict classifications across a dataset of 7.1 million sentences. This approach generates key variables at the patent level, including whether a patent contains at least one of these expressions ("occurrence") and the proportion of text devoted to these expressions ("density"). We correlate these expressions with a patent's technological orientation, which includes two broad categories: socially oriented or otherwise. A socially oriented patent is one that, based on its objective technological features, is better positioned to address societal issues. This variable is operationalized through mapping to UN Sustainable Development Goals (SDGs) based on technological classification provided by LexisNexis Intellectual Property Solutions. Our findings reveal a significant alignment between narrative promises and technological focus. PVEs show a positive association with social orientation - patents mapped to SDGs based on their core technical attributes. Conversely, PRIVEs exhibit a negative correlation. Most notably, PVE-PRIVEs demonstrate the strongest positive association with social orientation, suggesting that patents articulating both public and private values may be best positioned to address societal challenges while maintaining commercial viability. This aligns with Porter and Kramer's (2011) concept of shared value creation. These results challenge the notion that VEs in patents are merely rhetorical tools, contrasting with other value-based statements in science, technology and innovation (STI) that have been criticized for lack of enforceability and potential for "ethics-washing" (Munn, 2023). The correlation between PVEs and social orientation suggests these expressions may reflect or indicate an invention's genuine potential for societal impact. The research contributes to discussions about value-based statements in STI, including soft law tools like guidelines and standards to govern emerging technologies (Schiff et al., 2020), broader impact statements in research funding (Bozeman and Youtie, 2017), and Ethical, Legal, and Social Implications (ELSI) within technology initiatives (Fisher, 2019). We propose that VEs may play an important role in responsible innovation frameworks and anticipatory governance by articulating potential societal impacts ex-ante (Guston, 2014). This study responds to recent calls to develop "new quantitative methods to identify the public value of patents" and "identify the societal value or harm of intellectual property rights in a systematic way" (Castaldi, 2024). By demonstrating meaningful alignment between narrative promises and technological orientation in patents, we contribute to understanding how value-based statements in STI may substantively reflect and potentially influence the development of socially beneficial technologies. References Bozeman, B., & Youtie, J. (2017). Socio-economic impacts and public value of government-funded research: Lessons from four US National Science Foundation initiatives. Research Policy, 46(8), 1387-1398. Castaldi, C., Giuliani, E., Kyle, M., & Nuvolari, A. (2024). Are intellectual property rights working for society?. Research Policy, 53(2), 104936. Fisher, E. (2019). Governing with ambivalence: The tentative origins of socio-technical integration. Research Policy, 48(5), 1138-1149. Goos, M., & Savona, M. (2024). The governance of artificial intelligence: Harnessing opportunities and mitigating challenges. Research Policy, 53(3), 104928. Guston, D. H. (2014). Understanding 'anticipatory governance'. Social Studies of Science, 44(2), 218-242. Munn, L. (2023). The uselessness of AI ethics. AI and Ethics, 3(3), 869-877. OECD (2024). OECD Digital Economy Outlook 2024 (Volume 1): Embracing the Technology Frontier. OECD Publishing, Paris. https://doi.org/10.1787/a1689dc5-en Pelaez, S., Verma, G., Ribeiro, B., & Shapira, P. (2024). Large-scale text analysis using generative language models: A case study in discovering public value expressions in AI patents. Quantitative Science Studies, 5(1), 153-169. Porter, M. E., & Kramer, M. R. (2011). The big idea: Creating shared value. Harvard Business Review, 89(1-2), 62-77. Roco, M. C. (2023). National Nanotechnology Initiative at 20 years: enabling new horizons. Journal of Nanoparticle Research, 25(10), 197. Schiff, D., Biddle, J., Borenstein, J., & Laas, K. (2020). What's next for AI ethics, policy, and governance? A global overview. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (pp. 153-158).

ABSTRACT. The recent expansion of discussions on the need for more transparent practices in prioritizing, selecting, and funding research includes, among other factors, a closer examination of conventional evaluation procedures in science, particularly peer review. In this context, debates frequently arise regarding the strengths and limitations of this traditional method, which, despite being deeply institutionalized within scientific practice, remains susceptible to various biases and inaccuracies. Although discussions on this topic have intensified in recent years, empirical and exploratory research on peer review, especially in scientific journals or funding agencies, remains limited. This gap is largely attributed to the scarcity of available data, particularly in developing countries. This study, adopting an empirical and exploratory approach, aims to analyze the peer review process in project selection conducted by the São Paulo Research Foundation (FAPESP), a major research funding agency in Brazil. Specifically, it investigates the profiles of researchers and reviewers and their influence in selection processes. The analysis encompasses 37,423 research grant applications submitted to FAPESP between 2000 and 2017, covering both approved and rejected proposals. The database, containing information on the application processes, proponents, reviewers, and decision outcomes, was provided by FAPESP and complemented with publication data from CV Lattes, a public platform hosting the academic résumés of Brazilian researchers. Two complementary methodological approaches were employed for data analysis. The first approach applied a Baeyesian logistic regression to model the probability of grant approval in the programs studied, focusing on understanding variations in the distribution of profiles between approved researchers and those who submitted proposals. This analysis considered aspects such as field of knowledge, gender, seniority, and prior experience in obtaining research grants from FAPESP. To enhance the accuracy of the analysis, control variables were included: the field of knowledge of the proposal, the submission year, and the year of PhD completion of the proposer. These variables aimed to isolate the effects of potential confounders, ensuring that the relationship observed between proposal characteristics (and proponents) and funding decisions was accurately interpreted. This analysis allowed for the identification of funding patterns that may suggest the presence of systematic biases or discrepancies in resource distribution—a particularly relevant issue in the context of promoting equity and diversity in science. Approval rates in the analyzed programs ranged from 40% to 60%, with higher rates in the fields of biological sciences, exact sciences, humanities, engineering, and linguistics, and lower rates in agrarian, health, applied social sciences, and interdisciplinary fields. Interdisciplinary research, although categorized within traditional knowledge areas, generally had slightly lower approval rates than non-interdisciplinary research within the same areas, except in agrarian and health sciences. Regarding gender, female applicants generally had lower approval rates than male applicants, except in Applied Social Sciences, where women had a higher success rate. This gender gap tended to narrow after adjusting for the year of proposal submission and the year of PhD completion. Additionally, the proximity to the year of PhD completion was found to negatively impact funding approval chances. Each additional year of academic experience (measured from the year of PhD completion) increased the likelihood of proposal approval by 1%, highlighting the importance of seniority in funding decisions. The study also investigated the "Matthew Effect" in submitted proposals, examining whether those who had previously received funding were more likely to secure new grants. The findings suggest that applicants who had received FAPESP grants during the study period had an 18% higher approval rate than those without prior funding approvals. The second methodological approach utilized XGBoost, a gradient boosting machine (GBM) algorithm, to estimate classification probabilities, similar to the estimation of propensity scores. This analysis added reviewer profiles and proponents' scholarly output to the initial model, examining how differences in field of knowledge, gender, and seniority between reviewers and proposers might influence funding decisions. Additionally, the role of scholarly output, alongside other variables, was evaluated in relation to funding approval. In terms of influential factors, the field of knowledge emerged as the most critical determinant of funding approval, followed by the quality (measured by the SCImago Journal Rank, SJR) and quantity of published papers, seniority differences between reviewers and applicants, and the applicant’s PhD completion year. Gender and the match between the applicant's and reviewer’s fields were less influential. The relationship between these factors also revealed notable patterns. Approval rates were higher for researchers with greater scholarly output, both in terms of quantity and quality, and for more senior researchers, though the strength of these associations varied by field. For instance, in applied social sciences, the link between grant approval and publications was less pronounced than in biological and exact sciences. Conversely, seniority was a more significant factor for approval in fields like humanities and arts than in biological and exact sciences. Finally, a modest effect of the difference in academic age (measured from the PhD completion year) between reviewers and applicants was observed. Larger age gaps (with reviewers being more senior) correlated with lower approval rates, with distinct patterns across knowledge areas. Across these findings, the influence of applicant and reviewer gender, as well as their combinations, was minimal. The analysis highlights the existence of modest biases in the review processes of the studied funding agency, echoing both longstanding and recent criticisms of peer review. This study's findings contribute to broader discussions on research funding and evaluation practices, emphasizing the need for inclusive and diverse strategies to foster a more representative science capable of addressing current research challenges, especially in developing countries like Brazil. Such strategies are essential not only to promote fairness and equity within the scientific community but also to support a broader spectrum of innovative, high-quality research that responds to emerging social and scientific needs.

ABSTRACT. Introduction The capability of universities to place PhD graduates as professors at other universities is distributed unevenly. Network studies across various fields have confirmed this for the US university system, with Gini coefficients ranging from .62 to .76 [1-4], highlighting considerable institutional stratification. However, no comparable studies exist for German universities. We examine the exchange of PhD graduates between departments, using psychology and political science as examples for which comparative data are available. The German university system is an intriguing case: it has been argued that its institutional configuration, particularly the Humboldtian tradition, fosters a more egalitarian spirit than US higher education. For instance, professorial salaries in Germany are highly regulated under civil service laws, resulting in less salary variation than in the US [5]. Based on this, one would expect lower levels of both stratification and placement inequality. Conversely, the Humboldtian ideal of uniting research and teaching has been demystified as “invented tradition” [6]. Additionally, Germany’s excellence initiative (2005-2017) has recently intensified longstanding inequalities between large universities and smaller, less prestigious ones [7]. This suggests that the stratification and placement inequality may be closer to levels seen in the US. Theory A prominent network study by Burris [2] develops a prestige theory for the academic system, drawing on concepts from Max Weber (caste system) and Pierre Bourdieu (habitus). It posits that prestige is a structural property that organizes individuals and institutions hierarchically within a field. Prestigious departments possess specific characteristics, skills, and attitudes that afford access to desirable resources and tend to remain socially closed to protect their status. H1: Prestigious departments in German universities tend to recruit members of their professoriate predominantly from other prestigious departments, and thus among themselves. Less prestigious departments aspire to gain prestige. Since they cannot equip their graduate students with the same habitus as prestigious departments, they try to hire graduate students from prestigious departments to achieve a gain in prestige. H2: Less prestigious departments in German universities tend to recruit members of their professoriate predominantly from prestigious departments. Data and Methods This analysis uses data collected by a team led by Mark Lutter in 2019 and 2019, covering academic staff in psychology [8] and political science [9] at German universities. Data was collected from departmental and individual researchers’ websites and include information on their doctoral alma mater and current university. This resulted in datasets of 495 psychology professors across 74 departments and 279 political science professors across 73 departments. This study employs network analysis, treating departments as actors and professor placements as connection within the network. A connection exists between departments A and B if a doctoral student from A becomes a professor at B. Each department’s total professors trained is recorded as out-degree, and total professors hired as the in-degree. Departments are sorted by out-degree in descending order to determine their hierarchical position in the network. We test H1 and H2 by analyzing hiring movements within the placement network. For that purpose, we divide departments by their out-degree into quartiles. The first quartile contains 25% of departments with the lowest number of placements, the fourth quartile contains the 25% of departments with the highest number of placements. Results and relevance The Gini coefficient for out-degree is .61 for psychology and .66 for political science. This means that the capability of departments to train future professors is distributed similarly unequally in Germany as it is in the USA. Lorenz curves show that in psychology, 10% of departments have trained 53% of all professors. In political science, 10% of departments have trained 60% of professors. In psychology, 77% of professors within quartile 4 were trained there, and in political science, this proportion is 79%, supporting H1. For quartile 1, only 1% of psychology professors and none of the political science professors were trained there, lending support to H2. The training and hiring network of departments can be seen as an important mechanism for the emergence and dissemination of new ideas. The diversity of ideas largely depends on the diversity of doctoral training background [1, 3]. An excessive concentration on a few training institutions could thus pose an obstacle to the generation and spread of new ideas. The investigation of political science and psychology is a starting point for uncovering the hiring structures of the German university system. Clearly, two fields are not enough to make statements about the entire system. To address this, additional data sets are being developed to include sociology, economics, law, and geography. Data analysis is expected to be completed by spring 2025, enabling us to present broader and more in-depth findings on the entire system at the Atlanta 2025 conference. Reference list 1. Barnett GA, Danowski JA, Feeley TH, Stalker J. Measuring quality in communication doctoral education using network analysis of faculty-hiring patterns. Journal of Communication. 2010;60(2):388-411. 2. Burris V. The academic caste system: Prestige hierarchies in PhD exchange networks. American sociological review. 2004;69(2):239-64. 3. Katz DM, Gubler JR, Zelner J, Bommarito MJ. Reproduction of Hierarchy-A Social Network Analysis of the American Law Professoriate. J Legal Educ. 2011;61:76. 4. Wapman KH, Zhang S, Clauset A, Larremore DB. Quantifying hierarchy and dynamics in US faculty hiring and retention. Nature. 2022;610(7930):120-7. 5. Altbach PG. Paying the professoriate: A global comparison of compensation and contracts: Routledge; 2012. 6. Paletschek S. The invention of Humboldt and the impact of national socialism:: the German university idea in the first half of the twentieth century. 2001. 7. Heinze T, Habicht IM, Eberhardt P, Tunger D. Field size as a predictor of" excellence." The selection of subject fields in Germany's Excellence Initiative. bioRxiv. 2024:2024.03. 06.583816. 8. Lutter M, Habicht IM, Schröder M. Gender differences in the determinants of becoming a professor in Germany. An event history analysis of academic psychologists from 1980 to 2019. Research Policy. 2022;51(6):104506. 9. Habicht IM, Lutter M, Schröder M. How human capital, universities of excellence, third party funding, mobility and gender explain productivity in German political science. Scientometrics. 2021;126:9649-75.

ABSTRACT. Introduction and research questions

Scientific knowledge offers essential support for policymaking, serving as a valuable source of robust and reliable evidence to inform decisions. Research can be particularly useful for policymakers in navigating complex societal, environmental and economic challenges, such as the climate emergency, inequality, and public health (Capasso et al, 2019; Holtz et al, 2020; Viola et al, 2021). Despite its potential, accessing, and effectively using research remains a persistent challenge.

Considerable scholarly attention has been devoted to understanding the conditions, practices, and processes that facilitate the use of research-based knowledge in policy (Oliver et al, 2014; Cash et al, 2003; Bozeman and Sarewitz, 2005; Molas-Gallart and Tang, 2011; van der Arend, 2014; Gluckman and Wilsdon, 2016; Donovan 2011; Penfield et al. 2013; Thomas et al. 2020; de Rijcke et al. 2016; Watermeyer 2016). This focus spans multiple disciplines, with specialized journals such as Research Policy, Evidence & Policy, Environmental Science and Policy, Policy and Society, Health Research Policy and Systems, showcasing research use in and impact on policy as a core theme. Articles on the use of research in policymaking appear in high-level general journals such as Nature and Palgrave Communications (Schleussner et al., 2016; Donnelly et al., 2018; Kenny et al., 2017; Duncan, 2020).

There is a consistent interest from the research community in improving the uptake of research in policy. However, a byproduct of this sustained engagement from across fields and disciplines is a fragmented research landscape. Studies on research use, uptake, and policy impact emerge from diverse disciplines, each examining its contribution to policymaking, resulting in a wide-ranging, multidisciplinary body of work on the topic. It ranges from impact assessment into 'research utilization' into the much broader political science literature on the relationship between science and policy into much more applied studies on evidence-informed healthcare, as well as a body of literature on technology assessment and deploying research to address climate change and other environmental challenges.

In this study, we aim to understand this rich, but fragmented research landscape by identifying the key contributions and their intellectual foundations. We answer the following questions: how do different fields and disciplines approach research use in policy? What are the major theoretical and conceptual frameworks in the different niches of this thematic landscape? In other words, this study traces the use of research in informing and shaping policy decisions and changes.

We draw on the established methods for the analysis of the intellectual structure of research topics and fields: a combination of co-citation analysis and bibliographic coupling.

Method and Approach

This paper uses a keyword-based approach to identify relevant contributions across fields and disciplines that discuss the use and application of research in policy. Although dedicated journals exist that include research use and impact in their scope, our interest is in capturing broad discussions across domains.

As the first step, we developed a pilot study to test the approach. At the time of submission, the original approach was in the process of modification. Below, we outline the key steps and findings from the pilot study, noting the steps currently taken to improve the quality of the analysis.

Dataset Construction

This study was conducted as a part of a large-scale research project focusing on the impact of research, including a strand focusing on the role of research users for policy. Several literature reviews were conducted by the research team as a part of this work, which also highlighted the need for a more organised understanding of the literature landscape.

The keyword-based query for the pilot study was developed by the research team on the basis of the literature reviews (Edler et al., 2020; Karaulova and Edler, 2024; Edler et al., in review). It can be accessed in Table 1. The query was run in the Web of Science database in 2019 using the “topic” field. The search identified over 4,000 results, which were analysed. The research team was able to meaningfully interpret the results and made the decision to proceed with the analysis.

Table 1 Search Query

((research OR knowledge) AND (translation OR utilization)) OR

(("evidence synthesis" OR "evidence review" OR "evidence assessment") AND policy*) OR

((evidence OR research) AND (based OR informed) AND policy*) OR

((evidence OR research OR scien*) AND policy*) OR

("research impact" AND policy*) OR

("impact of" AND (research OR science)) OR

((science OR knowledge OR research) AND "into policy") OR

(("public value mapping" OR "contribution analysis" OR "productive interactions" OR ASIRPA) AND policy*)

The search query will be enriched by checking it against a dataset built as a part of a systematic review of the literature on research use in public sector organisations, including policymaking and healthcare settings.

The analysis will help understand the research landscape, how the different literatures are connected and whether they rely on the common or different knowledge bases. To achieve this, co-citation analysis and bibliographic coupling will be conducted.

Bibliographic coupling links reflect thematic similarities between publications based on the extent to which their bibliographies overlap. By looking at the bibliographic coupling map, we can examine cognitive proximities among the sources in the dataset.

Preliminary results

Despite our initial impression of a fragmented landscape, the literature discussing the relationship between research and policy clusters around key papers providing the key theoretical concepts and constructs, as revealed by the results of bibliographic coupling.

On the other hand, the results of the co-citation analysis suggest that beyond these shared foundational theories, the discussions around research use in policy are field- and discipline specific, suggesting that researchers engage in discipline-bound discussions, but will not necessarily be aware of the work being conducted on research use in policy in other research fields. The results contain many empirical papers that examine attitudes of scientists and policymakers, knowledge transfer strategies and barriers. These papers have overlapping findings, but do not necessarily cite each other. Although there is value in field-specific discussions, the lack of cross-field awareness of research progress may lead to double work and parallel discoveries, as well as to miss out on learning.

ABSTRACT. Research question: Gender bias in science has been extensively studied, and careers are a good example. Quite some research has focused on gender differences in becoming full professor, but other aspects of the academic career are less researched, such as gender differences in leaving academia (‘the leaky pipeline’ as this is often called). Here we focus on the question whether the gender difference in leaving academia has changed over time.

Data: We use two cohorts of PhD students from the same university, in order to have as little as possible variation between the two cohorts: The first cohort consists of all PhD students who received their degree between 2000 and 2006 (N=960), and we follow their career in the first seven years after graduation. The second cohort consists of all PhD students who received the degree between 2010 and 2014 (N=1000) and are also observed over a seven years’ period. As far as science is a merit driven system, one would expect that academic performance plays an important role in career advancement and in ending the academic career. However, performance has various dimensions, which may play a different role in different phases of the career. We assume that in the early career, research performance is the most important factor to remain in the system. As indicators for academic performance, we use bibliometric data – even when these are covering performance at best partly – are for the time being these are the best available. We excluded the fields where journal articles are not the dominant medium. Ignoring performance variables makes it impossible to distinguish gender differences from gender disadvantages or gender bias. Recent career studies therefore tend to include merit variables and mostly bibliometric data are used to measure research achievements. We include also variables measuring prestigious grants, and other prestigious signs of recognition, such as the cum laude award for the PhD thesis. Apart from merit, also other variables can be expected to influence the career, such as (i) personal characteristics like career preferences, which may be influenced by e.g. self-perception or gender stereotypes, (ii) family obligations which may lead to less time to invest in academic performance and to career breaks, and (iii) contextual factors such as the discipline, the reputation of the organization, and the PhD year (with the growth of the number of PhD students, a larger share of them may leave academia). The following data are collected about these early career researchers: (i) bibliometric performance variables, (ii) acquired prestigious grants, (iii) PhD with a honors recognition, (iv) the year of the PhD, (v) the faculty, (vi) research field, and (vii) career breaks as indicator for family obligations.

Methods: A Kaplan-Meier analysis will show whether women indeed leave academia more often and earlier than men do. An event-history analysis will be deployed to predict leaving the academic system, using the above-mentioned covariates. If gender has an effect on leaving academia after controlling for performance and other relevant covariates, one may conclude that the leaky pipeline is biased. Both analyses are done for the two cohorts, and then a comparison between the two cohorts shows whether the problem of the leaky pipeline is declining or not. Finally, for both cohorts we will try to create matched pairs, to have an experimental test of the effect of gender on the probability to leave academia.

Results: The analysis of the first cohort has been done, resulting in a statistically significant larger share of women – compared to men - is leaving academia. When including the covariates, gender does have a significant effect on leaving academia (women do it more often). The covariates have the expected effects: the performance variables correlate positively with remaining in the system, and the more recent PhD receivers in the first cohort have a higher probability to leave academia, suggesting that competition increases over time. The analysis of the second wave and a comparison with the first wave leads to the following findings: (i) In both the first cohort (2000-2005) and the second (2010-2014), men are leaving academia less frequently than women do. For example, in cohort 1, four years after the PhD, about 92% of the male sample is still publishing, whereas that only counts for about 84% of the women. For the second cohort the corresponding numbers are 89% and 80%. (ii) In the second cohort the pipeline leaks more than in the first cohort and this holds for women and men. With the increase of the number of PhD recipients (in cohort 1 about 75 per year, in cohort 2 about 110 per year), the retention rates have gone down. (iii) In the first cohort the gender difference increases stronger in the first cohort, and in the second cohort it is the other way around. So the early gender difference in retention has somewhat decreased, but in the later years the difference increases faster in the second cohort. (iv) Over the first seven years, the early cohort shows that the retention of men is 10% higher than of women, whereas in the later cohort the difference is 15%. So the gender difference increases over time, despite policies aiming at gender equality

Conclusion: Gender bias (in favor of men) in leaving academia seems to increase

Policy relevance: According to many observers, the working environment within universities has deteriorated, and the competition has intensified, making the universities a less attractive place to work and for a career. Women would suffer from this more than men do, and that may lead to women more than men leaving the university. This study is one way of investigating this problem: are women indeed increasingly leaving academia, and stronger than men? Of course, follow up research has to go deeper into the possible causes of the leaky pipeline, such as performance differences, different career preferences, or gender bias. An analysis of the (changing) patterns is a first necessary step.

ABSTRACT. Background: The accelerating impacts of climate change demand a rapid and sustainable transition from fossil-fuel-based energy systems to renewable alternatives. In this context, hydrogen gas is gaining a lot of attention, and South Korea has shown a strong commitment to hydrogen, pursuing various policy initiatives and R&D efforts from an early stage. For instance, the South Korean government announced national energy strategies centered on hydrogen energy in 2005 and 2019, and these strategies reflect a commitment to increasing hydrogen use across society, not just in fuel cells (Kim et al. 2024). However, contrary to traditional views that start-ups primarily drive energy transitions, public fossil fuel enterprises, based in natural gas, are attempting to lead this transition in Korea. Korea Gas Corporation has set a 2030 vision as “Everywhere Green Life, H2 KOGAS,” and similar agencies are advancing hydrogen safety standards and hydrogen plant technologies. This study situates this case within the context of incumbent-led sustainable energy transitions, using the Multi-Level Perspective (MLP) framework. MLP studies have argued that innovations develop within niches driven by new entrants (Kemp et al. 1998; Smith and Raven 2012). Although a growing body of research examines incumbent actors' roles, the mechanisms remain underexplored (Berggren et al. 2015; Turnheim and Geels 2019). This study specifically examines how incumbents consolidate multiple niche innovations, leveraging technological knowledge, institutions, policies, and infrastructure. By integrating transition studies with organizational theory, it investigates the evolving strategies and roles of South Korean public enterprises in advancing hydrogen energy transitions (van Mossel et al. 2018; Mäkitie 2020).

Methods: This research conducted a qualitative multiple-case analysis of three South Korean public natural gas enterprises through the MLP framework. Data were collected from news articles, internal publications, and annual reports spanning the 1990s to early 2020s. The study also analyzed enterprises’ intellectual property portfolios, internal research projects, and infrastructure assets by using public data. I used public data and the WIPO database to cover all patents filed by public gas enterprises for the patents (n=768) and selected hydrogen-related patents among them.

Findings: Using the MLP framework, this research revealed strategic and organizational shifts among incumbent regime actors over the past two decades. First, in terms of landscape-regime interaction, obligations from international agreements intensified as South Korea’s economy grew, increasing pressure on public enterprises to lower carbon emissions. Second, as public enterprises had built expertise in gas infrastructure, they were able to independently conduct R&D and patent hydrogen-related technologies at the niche level. Lastly, national hydrogen policies evolved alongside the maturing natural gas regime, enabling the repurposing of existing infrastructure, such as pipelines, for hydrogen production and transport.

Significance: These findings enhance our understanding of how public enterprises adapt to climate challenges, emphasizing the role of technological infrastructure as a long-term asset. Moreover, the research offers insights for policymakers and public enterprises, particularly in resource-limited countries, on leveraging existing assets for sustainable energy transitions.

ABSTRACT. In the 1970s, the slum upgrading approach was embraced by the Philippine government in response to the shelter dilemma facing the urban poor. However, rapid urbanization has led to the proliferation of slum areas in major cities, including Quezon City. One of the significant slum areas in the city is found in Barangay Kamuning, specifically the Bernardo Park area where this study is focused on. Meanwhile, cities nowadays including slum settlements are based on linear metabolisms contributing to different impacts such as the high dependency on non-renewable ones. In this context, this paper aims to apply the Extended Urban Metabolism & Sustainability Indicators to seek spatial planning solutions to improve living conditions of dwellers in Barangay Kamuning’s slum settlement.

The sequential explanatory approach was used in the study with quantitative approach (i.e. Life Cycle Assessment) being the first stage of data collection then supported by the qualitative approach (i.e. in-depth semi structured interviews with stakeholders to make the study as participatory as possible). The gathered data was used in generating plans to be applied elsewhere in the city or in other slum settlements with similar landscape characteristics. The study's findings reveal the environmental impact of Quezon City's metabolism, with depletion of abiotic resources, particularly fossil fuels, was identified as the most significant impact category. The identified categories then served as Sustainability Indicators, enabling targeted interventions to enhance the city's environmental sustainability with the aim of improving the liveability conditions of slum dwellers. The study's assessment provides valuable insights for policymakers, urban planners, and relevant stakeholders in prioritizing efforts towards a more sustainable city metabolism in slum upgrading.

ABSTRACT. This study investigates the conceptions of climate change held by Indonesian upper-secondary teachers and their willingness to implement climate change education (CCE) in the classroom. Using a mixed-methods approach with 329 teacher respondents, this research examines the complex interplay between cognitive, affective, and behavioural dimensions and their role in influencing teachers’ engagement in CCE. Structural equation modelling analysis revealed that a positive attitude towards climate change is the strongest predictor of willingness to act, with self-efficacy and situational support also playing critical roles. Findings highlight the importance of both personal and contextual factors: teachers’ confidence in their knowledge, support from school environments, and the affective impact of concern and mixed hope towards climate change all contribute to shaping their attitudes and willingness to implement CCE. Notably, perceived knowledge rather than actual knowledge was found to have a significant impact on teachers’ self-efficacy and willingness to act, underlining the knowledge-action gap that exists within this context. Additionally, widespread misconceptions persist among teachers, with many incorrectly identifying the ozone hole as a primary cause of climate change and overestimating the effectiveness of individual actions like waste recycling in mitigating emissions. This underscores the need for targeted, scientifically accurate training and resources to address these gaps. The study also identifies emotional responses among teachers—characterised by both concern and limited optimism—as factors that complicate CCE implementation. This research highlights the necessity of fostering supportive environments that enhance teachers’ self-efficacy and provide resources and pathways for effective climate change instruction. Insights from this study can inform policymakers and educational leaders in developing curricula and professional development programmes that enhance teacher engagement in CCE. Ultimately, these findings contribute to the global effort of equipping educators to build a more sustainable future for students and communities.

ABSTRACT. Lucio Flores (3) Jorge Tello (1)

1.- Introduction

The accelerated technological advances and the exhaustive commercial competition at the international level have accelerated the pace of action of the States to support the productive sectors and make them more productive, in particular, through financing for the generation of science, support for technological development and incentives for innovation, considering that they are key factors for economic growth and being able to impact on social welfare. In Mexico, the Federal Government has joined these valuable efforts, implementing plans, programs and projects as part of a public policy that channels resources (economic, human and technical) towards the construction of capacities at the local level in science, technology and innovation (STI).

The above has generated a broad discussion and a rigorous analysis in order to understand that state capacities do not arise from the Nation-State, in a vision of central control, but that these capacities are built and tested in the territories through the presence of the State (national-federal) to contribute to the provision of public goods (such as the generation of knowledge), either because they are not the responsibility of the subnational entities themselves or because these entities do not have the resources to provide them.

From this perspective of provision through the intervention of the State, a debate arises that is based on what Stiglitz (1999) raised some time ago regarding what role does the State play in the provision of knowledge so that it is not insufficient? In this context, the Federal Government in Mexico has undertaken various actions through the body that coordinates the scientific-technological sector of the country, the now called National Council of Humanities, Science, Technology and Innovation (CONAHCYT) to finance part of the R&D in the country, support innovation processes, the training and consolidation of high-level human capital, and providing direct support so that various agents of the national STI system are linked with the purposes of innovating and improving products, services and processes that generate an impact on social well-being, without neglecting care for the environment.

2. Research questions

In the trajectory of the set of actions implemented by CONAHCYT, it is worth asking: what role is played by the distribution of resources (economic, human, material and technical) both existing and those distributed by the Federal Government in the construction of STI capacities at the state level? Which federal entities are taking the most advantage of decentralization and local capacity building strategies? Why is it important to advance in the construction of STI capabilities endogenously at the federal entity level in a global context like the current one?

In this sense, the research aims to carry out a multidimensional characterization of the importance of developing scientific-technological and innovation capabilities at the local level, based on the initial conditions and the actions executed by the federal executive through CONAHCYT, in such a way that, through an analysis of the main indicators, a better understanding of whether the implementation of the STI policy is positively and adequately promoting the construction of said capabilities at the state level is achieved.

3. Methodology

In order to approach the answers to these questions, it is necessary to make an evaluation of the resources currently available and to be able to make a geographic map of the progress in STI in the 32 federal entities.

To evaluate and show the location of the resources and how the STI capabilities are formed at the federal entity level, the methodology used is based on the quantitative paradigm. Deductive logic is put into practice and two multidimensional indices are designed: the first is a synthetic index and the other is based on the theory of fuzzy logic. The first is a tool that allows us to glimpse the current capacities at a local level and make an objective interpretation. The second is an analytical instrument that allows us to show the areas of opportunity and aspects to improve, allowing us to show the route to direct greater efforts towards capacity building. The contrast between this pair of indices allows for a better analysis given the amount of information available for the 32 federal entities, and thus, to make projections and better public policy recommendations.

For this analysis, various variables were selected in terms of their being susceptible to being observed and measured among them, directly related to the construction of capacities in STI. The selection of the variables revolves around three axes: training and strengthening of high-level human capital (including physical infrastructure), the financial resources that are allocated locally and through the programs implemented by CONAHCYT and the scientific and technological production as a result of the two previous ones.

4.- Some Findings

The results show that both the support and the capacities around the promotion of STI are an important part of the decentralization strategy, however, they are strongly localized in only some federal entities, showing not only a significant imbalance between them, but also the weakness of the strategy followed by the Federal Government for the construction of capacities at a local level.

With the synthetic index, high concentrations of economic, human and infrastructure resources are perceived in only some entities of the country, highlighted in all the programs implemented by CONAHCYT, that is, CDMX, Nuevo León, Jalisco, the State of Mexico and Puebla, which have important levels in the formation of capacities to promote STI, thanks to the absorption of public financing. On the other hand, with the diffuse index, CDMX, Jalisco, Nuevo León, the State of Mexico, Puebla, Veracruz, Baja California and Guanajuato are the entities with environments that range from acceptable to favorable in said construction.

In addition, the results show the need to propose new regulatory and institutional mechanisms to achieve firm agreements and associations between the public and private sectors on issues related to venture capital, fiscal incentives or others, which allow for the achievement of clear and common objectives, so that the contribution of STI is in favor of the economic development of the entire country.

ABSTRACT. The setting and achieving of renewable energy targets and climate goals are also faced with the inherent instability of energy production, consumption rates, and the ever-changing environment. The conventional approaches to management emphasize the setting of targets based on static case models that cannot take into consideration the actual real-time data, country resources, and more importantly, the dynamic climate conditions. This paper examines a case of an application of artificial intelligence-based predictive analytics in renewable energy and climate policymaking to bring more accurate and achievable plans and targets into consideration given that the drive towards renewable energy and climate targets is still in its infancy. The study addresses the following research questions: For what purpose may affective, accurate, and timely AI-based predictive models be utilized to evaluate the viability of renewable energy targets and emissions reduction goals? How does AI take into consideration each country’s energy production, consumption, and resources in setting achievable renewable energy goals? Where, then, lies the place of AI in achieving these changes and a dynamic adjustments of policies to meet energy targets based on changing economic, environmental, and technological conditions?

To answer these questions, we apply AI predictive modeling based on computing methods such as regression analysis, support vector machines, or time series analysis. The critical variables used in the analysis include countries, years, GWh energy generation, CO2 emissions, renewable energy goals, and natural resource prospects, including solar radiation, wind velocity, hydro energy, and biomass. These bits of information are used for developing prognosis models that assess the likelihood of achieving RE and emission reduction goals based on the country’s energy mix and endowment with natural resources.

According to the study, AI can significantly increase the efficacy of renewable energy targets when paired with real-time data on energy generation capacity and demand, as well as lessen environmental stress. The numerous AI models foretell the possibility of reaching particular renewable energy goals and compare the presented natural resources like hydro, wind, and solar energy with a country’s consumption and the emission reductions required. The figures also make it possible to monitor progress and set fresh objectives where necessary due to the real-time results generated by predictive analytics. For instance, in areas of high solar intensity, the AI models recommend increased reliance on solar power, while in areas with vast hydropower potential, the author’s AI blurbs for increased investment in the generation of hydropower.

This research can be beneficial for policymakers as it shows how it is possible to achieve those goals more efficiently, thus allowing the setting of more realistic targets. AI and prediction assist in guaranteeing that the set goals are achievable because they are based on resources’ availability and consumption, thereby enhancing efficiency in energy policy formulation. Furthermore, it reveals the potential and utilization of dynamic policy optimization approach to energy and climate targets and constraints, making an efficient and effective responsive energy policy.

ABSTRACT. Cities are launching various smart city projects to resolve their urbanization issues and challenges (Snow et al., 2016; Yin et al., 2015). Cities are adopting smart city projects and technologies to improve service delivery, sustainability, infrastructure, quality of life, and governance (Angelidou et al., 2018; Giffinger et al., 2007). As of 2015, about 250 smart city projects have been launched worldwide (Angelidou, 2015). Despite the proliferation of smart city projects, some are failing and scaling back on their promises (Sax, 2022). While we know why some smart cities are vibrant and thriving, we know less about why some projects face false starts or are abandoned. My paper attempts to address this gap in the literature. I intend to focus on how public-private partnerships and attitudes affect the success of smart city projects using the mixed method research approach, interviewing elites and surveying citizens in Accra, Ghana, and Kigali, Rwanda.

ABSTRACT. This paper explores the tension between competitive and openness norms underlying data-sharing practices in scientific research, focusing on scientists' dual demands for career advancement and social responsibility. As open science has increasingly become a core concept among academics and policymakers, data sharing is widely promoted for its potential to enhance research transparency, replicability, and interdisciplinary collaboration (Vicente-Saez & Martinez-Fuentes, 2018; Wilkinson et al., 2016). However, scientists face multiple competitive pressures in practice: while the transparency and efficient resource promoted and strongly supported by open science, the demands of academic competition and intellectual property protection present significant barriers (Borgman, 2012). Grounded in social exchange theory and norm conflict theory, this study addresses three key questions: (1) How do competitiveness norms inhibit scientists' data-sharing behavior? (2) How do openness norms encourage data sharing? (3) How do tensions between these norms affect scientists' data-sharing decisions under different contextual pressures? Social exchange theory provides a framework for understanding how scientists weigh personal benefits against data-sharing costs (Haeussler, 2011), while norm conflict theory explains the decision-making process when scientists face competing pressures from both competitive and openness norms (Franzoni & Sauermann, 2014). This study tests these hypotheses using data from the 2024 Science Impartiality Survey, a representative random sample of U.S. academic scientists across six disciplines, collected by the Science Panel Opinion Survey at Arizona State University. The variables in this study include scientists' data-sharing behavior (dependent variable), competitiveness norms and openness norms (independent variables), and data sensitivity and policy pressure (moderating variables). Competitiveness norms primarily manifest as scientists' concerns for data confidentiality and intellectual property protection in the face of publication pressure and career advancement needs, while openness norms encompass academic transparency needs and social responsibility, emphasizing data openness and the social contribution of scientific research. As one moderating variable, data sensitivity reflects the degree of privacy, ethical, or commercial risk associated with scientists' data, and policy pressure. Another moderating factor, policy pressure, comes from requirements set by funding agencies and academic journals for data sharing. By exploring the tensions scientists face in open science practices, this study enhances understanding of the motivations behind scientists’ data-sharing practices. It also offers policy recommendations to balance incentives for data sharing with career advancement needs, promote sustainable open science, and address public expectations and concerns.

Reference Borgman, C. L. (2012). The conundrum of sharing research data. Journal of the American Society for Information Science and Technology, 63(6), 1059–1078. Franzoni, C., & Sauermann, H. (2014). Crowd science: The organization of scientific research in open collaborative projects. Research Policy, 43(1), 1-20. Haeussler, C. (2011). Information-sharing in academia and the industry: A comparative study. Research Policy, 40(1), 105–122. Vicente-Saez, R., & Martinez-Fuentes, C. (2018). Open science now: A systematic literature review for an integrated definition. Journal of Business Research, 88, 428–436. Wilkinson, M. D., Dumontier, M., Aalbersberg, I. J., Appleton, G., Axton, M., Baak, A., ... & Mons, B. (2016). The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data, 3, 160018.

ABSTRACT. As artificial intelligence (AI) progresses, its potential to reshape labour markets has become a topic of growing scientific interest. While some studies suggest that AI has the potential to lead to productivity gains throughout the economy (Agrawal et al., 2019), there are also fears about AI’s potential to replace human labour (Frey & Osborne, 2017) and that its transformative effects may be somewhat limited (Acemoglu & Restrepo, 2019). Although early evidence suggests that AI positively contributes to labour productivity (Li & Raymond, 2023; Noy & Zhang, 2023), the pervasive and general-purpose nature of AI calls for further research to fully understand its impact on work and workers. It especially remains crucial to uncover its implications within the workplace to better understand the mechanisms of how AI changes workers’ productivity.

This paper presents the empirical evidence on the effect of AI on workers’ productivity by exploiting a field experiment in a large European financial service provider in Europe. By leveraging the firm’s key performance indicators (KPI), we study the introduction of a labour-augmenting AI tool for the training regime of its call centre agents. The data contains detailed information on agents’ productivity ranging from simple quantitative measures—e.g., call handling time—to complex AI-generated measures of the agents’ communication style. We use these metrics to estimate the effects of AI-enhanced training on agent’s productivity.

Following the firm's staggered introduction of the AI tool, we find that AI-enhanced training significantly improved worker productivity by reducing agents’ average call handling time (AHT) by approximately 14%. The reduction in AHT was mainly driven by a decrease in hold times and speaking time. We find that AI implementation also improved the quality of agents’ communication through a reduction in usage of less-desired communication techniques, such as stop words and diminutives. The treatment was particularly effective in reducing longer calls. Our heterogeneity analysis shows that the productivity gains and improved communication styles were substantially stronger for low- than for high-tenured agents.

Overall, our paper contributes to the existing literature on the intersection of AI and labour productivity and highlights the potential for AI to act as a mediator in training processes. Our findings suggest that AI tools can be effectively integrated into training regimes to boost overall worker performance.

ABSTRACT. Artificial intelligence (AI) has entered the pacing stage of its technology cycle, marked by rapid advancements in performance and a sharp rise in patent filings since 2020. This stage is crucial—its future trajectory will significantly shape its societal and environmental impact. Patents, as one of the innovation indicators, offer insights into strategic areas of focus within the field and the future directions of technological progress. This study draws on two primary theories of patents. Reward theory posits that patents incentivize innovation by granting territorial exclusive rights, reflecting the national priorities of different countries. Contract theory highlights patents as instruments of social exchange: they provide temporary protection in return for the public disclosure of innovations, fostering broader societal benefits over time. While the societal implications of AI patents are increasingly important, the empirical evidence in this area remains under-explored. Focusing on AI patents from the U.S. and China, two global leaders in AI, this study explores how recent innovations align with the United Nations Sustainable Development Goals (SDG) to evaluate their potential societal and environmental impact.

This study identified AI patents using keywords defined in the World Intellectual Property Organization (WIPO) report, including terms such as machine learning, neural networks, and deep learning, alongside related Cooperative Patent Classification (CPC) codes. These patents were searched and retrieved from The Lens, covering granted applications from January 2020 to June 2024. After removing duplicates, a total of 69,523 patents from the U.S. and 126,485 patents from China were identified. Textual descriptions of these patents were then extracted using Google Patents. To assess the SDG foci of the patents, a high-performance pre-trained large language model (LLM), Llama 3.1, was used to compute the cosine similarity between patent descriptions and SDG keyword embeddings. Similarity scores were evaluated to classify patents as relevant to an SDG based on thresholds determined through iterative sampling and manual review.

The results showed that, among all SDGs, ‘SDG3—Good Health and Well-being’ was the dominant focus for both the U.S. and China. Several SDGs, notably ‘SDG5—Gender Equality’ and ‘SDG14—Life Below Water,’ received significantly less attention. The highest number of patents were related to economic SDGs, followed by environmental and social SDGs. A chi-squared analysis of patent distribution across SDGs between the two countries showed significant results. Post hoc analysis indicated that China contributed more patents to environmental SDGs, particularly ‘SDG7—Affordable and Clean Energy’ and ‘SDG12—Responsible Consumption and Production,’ while the U.S. led in ‘SDG3—Good Health and Well-being,’ ‘SDG4—Quality Education,’ and ‘SDG10—Reduced Inequalities.’ Although the U.S. had more unique AI patents related to SDGs overall, the gap between the two countries narrowed over the observed period.

These findings suggest a common underrepresentation of certain SDGs in AI patenting, particularly those addressing social equity. Notably, U.S. patents aligned more with economic sustainability, while China’s AI patents demonstrated strong potential for advancing green innovations. The results, with reference to reward and contract theories, highlight that while patent distribution can serve as early indicators of SDG progress, divergent national priorities point to an increasing reliance on certain countries to drive solutions to global challenges. This study highlights the importance of the need for well-coordinated international partnerships to harness the unique strengths of each nation in AI innovation toward achieving sustainable development.

ABSTRACT. Cities throughout the world are increasingly challenged to improve urban sustainability and resilience due to climate change impacts while experiencing transformative innovation. However, existing energy mix strategies trigger a tradeoff in urban innovation efforts in achieving these objectives that are central to smart city governance. While less carbon emissions require greater reliance on renewable energy sources in the energy mix, it is exposed to inherent vulnerabilities from varying weather conditions, often leading to large-scale power outages and blackouts during extreme weather events. Although emerging technological advances in the clean energy industry could address many of these conundrums, operational risk, and energy affordability must be ensured, and our conceptual understanding and empirical evidence of such developments remain largely absent from the current urban innovation and smart city literature.

This paper first documents these challenges using cross-country analysis of energy mix, extreme weather events, and resultant losses of various socio economic outcomes among major global cities in the world. It then explores small modular reactors (SMRs), an advanced form of clean nuclear energy that provides greater means for safety and climate resilience, compared to traditional nuclear power plants. We lay out conceptual foundations in which SMRs can play a key role in enhancing urban sustainability and resilience while gathering public support and trust in the operational safety of nuclear energy.

Focusing on the applicability of SMRs and their integration with renewables in the energy mix within the broader context of urban innovation literature, we present key opportunities, challenges, and risks in leveraging these technological advances for smart city development. By using natural language processing techniques, we analyze national energy transition plans, R&D spending, and policy development guideline documents. This paper will present a comparative analysis uncovering which countries are more prepared to become prominent leaders in this innovation space. We conclude the paper with empirical cases where cities are actively seeking to incorporate SMRs in their economic development, energy reliability, and climate resilience strategies.

The findings will inform how cities can invest towards creating more sustainable and climate-resilient urban environments as well as more robust energy security through the lens of energy mix and technological perspectives.

ABSTRACT. Research question The rapid development of digital intelligence technology has brought multiple opportunities and challenges to social governance. As an important part of the country's efforts to modernize its governance system and capacity, community governance in China is accelerating its digital intelligence transformation: some successfully revitalize their communities, while others deviate from their original intention and become formalism. The fundamental difference lies in whether residents can co-create public value with the community during the digital intelligence transformation, as well as achieving value co-creation. The community emotion is a crucial factor in this process. Based on this phenomenon, the research question of this paper is: How does residents' community emotion affect value co-creation in the digital intelligence transformation of communities? Methodologies As an individual behavior of residents, the implementation of value co-creation requires residents to invest time and resources, which follows a certain decision-making logic rather than a simple behavioral inertia. The Theory of Planned Behavior (TPB) is widely used to explain individual behavior. As to the public's use behavior of technology, the Technology Acceptance Model (TAM) is often used. With the development of digital intelligence technology, the public and scholars realize that technology may bring risks as well as benefits, so "perceived technology benefit" and "perceived technology risk" have been increasingly introduced into the technology acceptance model. These two theories have a good explanatory power in explaining technology-enabled individual behavior. Therefore, this paper puts forward these hypotheses and theoretical model "community emotion - technological perception - value co-creation" (Figure 1) based on these two theories.

Figure 1 Hypotheses and theoretical model

Scenario simulation questionnaire experiment is used to answer the research question. According to the field research in Chinese communities, there are three main logics to utilize digital intelligence technology to promote residents' value co-creation: broaden residents' participation channels, use incentive mechanisms, and enrich residents' self-presentation opportunities. Therefore, three sets of scenarios are developed to simulate these three logics. The questionnaires are then distributed to Chinese citizens randomly through online link and finally 1,780 valid questionnaires are received. By constructing a structural equation model of residents' intention for value co-creation, this paper further explains how community emotion affects residents' intention for value co-creation in a technology-empowered context. The roles of community emotion and the effect of technology empowerment in different scenarios are compared and discussed to figure out a better solution of technology-empowerment. Findings First, the bigger the uncertainty brought about by technological empowerment, the stronger the need for community sentiment to help bring about a stable transformation. The more benefit perception and risk perception brought by technology empowerment, the greater the direct and total effect of community trust on value co-creation intention; Similarly, community interest in public affairs also plays a more significant role in technology benefit perception and risk perception. However, under the same risk perception condition, the importance of community interest in public affairs is greater than community trust. In other words, when the uncertainty brought by technology is bigger, the community emotion of residents plays an important role in "maintaining stability" : community trust makes residents willing to participate in technology-enabled value co-creation at high risks (community trust balances risk perception), and community interest in public affairs further stimulates the value co-creation opportunity brought by technological advantages (community interest in public affairs improves benefit perception). The most important thing to play the role of "maintaining stability" in community emotion is community public affairs interest, followed by community trust. In other words, residents' interest in community public affairs can make them effectively perceive the potential benefits and risks brought by technology empowerment, while residents' community trust is conducive to their belief in the realization of potential benefits and the avoidance of potential risks. Second, residents have the highest perception of benefits of the technology enabling incentive mechanism, and the effect of technology enabling self-presentation on promoting value co-creation is the worst. It can be seen that residents' perception of technological benefits from high to low is as follows: use incentive mechanisms, broaden residents' participation channels, enrich residents' self-presentation opportunities. Technology risk perception from high to low is: enrich residents' self-presentation opportunities, use incentive mechanisms, broaden residents' participation channels. However, only in the scenario of “broaden residents' participation”, residents' perception of technological benefits is significantly lower than in other scenarios. It can be seen that by using incentive mechanisms, residents can generally perceive that technology brings more benefits, and the incentive means can promote the value co-creation of residents, but the specific incentive effect is also related to the incentive intensity. Conclusions The main findings are as follows: First, in the context of the digital intelligence transformation of community governance, residents' community emotion remains a crucial factor in promoting residents' value co-creation. Second, community emotion that affects residents' value co-creation mainly includes community trust and interest in community public affairs, which influence residents' technological perception and further affect their intention for value co-creation. Third, under different technology-empowered scenarios, residents' community emotion has different influence mechanisms on their intention for value co-creation, and the effect of technology empowerment varies. Among them, residents have the highest perception of the benefits of technology-empowered incentive mechanisms, while technology empowerment to enhance self-presentation has the worst effect on promoting value co-creation intention. Traditional technological determinism believes that simply introducing digital intelligence technology can stimulate residents' value co-creation. However, this study finds that residents' community emotion is still the key factor influencing residents' value co-creation in the digital intelligence transformation, indicating that the government still needs to adhere to its original intention of humanism in promoting digital intelligence transformation. At the same time, this paper analyzes the internal mechanism of residents' value co-creation in technology-empowered scenarios from a behaviorism perspective, which provides some practical evidence for the application of value co-creation theory in the field of community governance.

ABSTRACT. The concept of entrepreneurial ecosystems (EEs) has received significant attention due to their crucial role in fostering entrepreneurship. Over the past few decades, a substantial body of literature has explored how interactions among key elements—such as governments, entrepreneurs, and research institutions—stimulate entrepreneurial activities within EEs. However, our understanding of how these interactions drive the development of EEs, particularly in shaping their resilience, remains incomplete. This research gap may impede a comprehensive understanding of the heterogeneities of EEs over time and across different regions. Our study aims to address this gap by providing a detailed classification of multi-entrepreneur interactions within EEs, elucidating how these interactions influence the functioning of EEs.

This research builds on the dynamic lifecycle model of entrepreneurial ecosystems proposed by Cantner et al. (2021) and its adapted variant by Shi et al. (2022), which offer a structural and chronological framework for examining the evolution of an EE. Drawing insights from their work, we conducted an in-depth case study of the Shenzhen entrepreneurial ecosystem (EE) from 1980 to 2020, investigating the interactions between two major types of entrepreneurs: policy entrepreneurs (from government) and business entrepreneurs (from the private sector). Policy entrepreneurs are further categorized into central and local government policy entrepreneurs, while business entrepreneurs are divided into leading companies and small- or medium-sized enterprises (SMEs).

This study employs the process-tracing method, utilizing archival and interview data to map the underlying interactions between different entrepreneurs during the development of the Shenzhen EE. Process-tracing has been widely used for modeling and assessing causal relationships or mechanisms in complex systems. Moreover, it emphasizes the sequence and timing of events, making it ideal for capturing multi-entrepreneur interactions over time. We propose that the evolution of the Shenzhen EE can be divided into four phases: pre-emerging, birth, growth, and maturity. Each phase is characterized by specific entrepreneurial activities or phenomena, carried out by distinct entrepreneurial stakeholders. Additionally, the evolution of the Shenzhen EE is shaped by historical reforms and regional cultures. Process-tracing enables us to further illuminate how interactive mechanisms can be influenced by different cultural and social contexts.

Our findings indicate that interactions between policy and business entrepreneurs over the past four decades have not only created diversity at the micro-level but also established coherence at the macro-level within the Shenzhen EE. This diversity and coherence are crucial "inputs" for building the EE's resilience. By examining the underlying interactions between different entrepreneurial agents, this research bridges the gap between innovation policy literature and entrepreneurship literature. This research provides valuable insights for stakeholders across various sectors, including policymakers, entrepreneurs, and researchers interested in the dynamics and resilience of entrepreneurial ecosystems.

References:

Cantner, U., Cunningham, J. A., Lehmann, E. E., & Menter, M. (2021). Entrepreneurial ecosystems: a dynamic lifecycle model. Small Business Economics, 57, 407-423.

Shi, X., & Shi, Y. (2022). Unpacking the process of resource allocation within an entrepreneurial ecosystem. Research Policy, 51(9), 104378.

ABSTRACT. Introduction We studied solar taxiing, an e-mobility initiative in Sub-Saharan Africa, to contribute to the discussions on responsible management (RM) practices by exploring new forms of arrangements through which responsible management practices emerge alongside new ventures. We investigated unique dynamics of how various actors assemble to frame and constitute responsible management practices, defined as the integration of ethics, sustainability and responsibility into routine activities to harmoniously fulfil stakeholder interests (Carroll et al., 2020; Laasch, Suddaby, Freeman, & Jamali, 2020b; Nonet, Kassel, & Meijs, 2016). Using the actor-network theory (ANT) approach, which is a socio-philosophical approach to understanding complexities by analysing relational elements (Arnaboldi & Spiller, 2011) to produce account of how science, technology and social realities are made (Latour, 2007), we investigated this initiative to study how actors assemble to translate climate change and social problems to responsible management practices. Grounded on ANT’s concept of agencement, which states that all participants in a system, be it humans or nonhumans, have equal agency to affect any given situation (Latour, 1987), we use the term “actors” to refer to both humans and nonhumans. Thus, we conducted this study to answer the following questions: (1) How do actors in socio-material entanglements assemble to create new form of arrangement that constitute responsible management (RM) practices and (2) how do responsible management (RM) practices emerge out of the socio-technical systems created through the assemblage process?

Methodology Data Gathering We gathered the empirical material using multiple sources including interviews and ethnographic observations. We conducted a 6-months on-site ethnography. The principal investigator travelled to Ghana for the fieldwork. He recorded his daily observations into the field notes, taking supporting pictures and videos. He conducted open-ended semi-formal interviews which lasted from about 15 minutes to an hour and 30 minutes. At the end of the ethnography 86 interview responses were gathered, along with 6 policy documents. Data analysis We unpacked the translations in the solar taxiing project by analysing four aspects of each event identified in the solar taxiing project. We focused on discovering – (1) what happened, (2) who and what were involved, (3) when did it happen and (4) how did it happen. These four thematic foci guided us to develop the thematic areas of the study. We used rounds of inductive and deductive coding to extract the order concepts, refine them, cluster them and aggregate them into themes and sub-themes. We used three rounds of open coding, axial coding and theoretical coding to develop a data structure using the Gioia methodology (Gioia & Pitre, 1990; Magnani & Gioia, 2023). We open-coded the interview transcripts and field notes by capturing key texts in the transcripts and field notes. In line with the Gioia methodology, after the open coding, we moved on to identify the first order concepts. We developed a third level of coding in which we fitted the codes into a coherent storyline.

Results and Findings We found that three dynamics occur in the translation process – (1) complementing and competing dynamics, (2) enabling dynamics, and (3) transformative dynamics. These complementary dynamics occur in the form of associations, partnerships, collaborations, appeals and commitments. Enabling dynamics occur through mandating, tasking, connecting, granting access and approving. Transformative dynamics involve actions such as implementing, capturing, assimilating, transforming, utilizing, testing and repurposing. It comprises roles conjointly performed by the actors to change a situation or to modify the identity of the actors. We found that these dynamics occur in four translation turning points of (1) applying old technologies in new ways, (2) enrolling new actors and developing new technologies, (3) (Mis)aligning interests and aligning abilities with opportunities, and (4) setting boundaries and objectives for the actor-network and protecting these objectives and boundaries with rules and regulations.

Discussions and Conclusion In this research, we studied solar taxiing, an e-mobility project which has been in operation in Sub-Saharan Africa since 2019, and we drew empirical insights to give an elaborate account of how various socio-material actors assemble in an actor-network, frame, and produce RM practices in unique ways that have not been explained by existing theories. Using actor-network theory as the heuristics and theoretical paradigm, we used the bottom-up approach rather than the top-down approach which has been widely established in extant literature as the way of understanding RM. This bottom-up approach helps understand how the empirical phenomenon connect to and deviate from existing theories (Caruana, Glozer, Crane, & McCabe, 2014). The approach to studying this was to investigate how localized RM practices emerge, get territorialized and institutionalized and disrupt existing models (Czarniawska & Joerges, 1996). Through the empirical analysis, we bring to fore the complexities and dynamics associated with the emergence of RM practices in a marginalized context, and in so doing, we open the gateway for discussions that challenges the status quo, the present structurisation of RM practices, and existing reporting frameworks. We advance that it is not always the case that there is consciousness and intentionality behind RM practices, but that RM practices could emerge unconsciously and unintentionally as a detour from some other mainstream intentional and conscious process. We demonstrate in this study that RM practices do not necessarily emerge from corporations. We found that organisations (particularly new ventures) could rather emerge as unintended product of RM practices. We theorise that (de)stabilising events in the form of (1) matters of concerns, (2) failures and successes, and (3) controversies and tensions result in various programs of actions which yield responsible management practices as both intended and unintended consequences, and that new ventures emerge out of these arrangements, quite unique from what we know in traditional literature that RM practices originate from corporations.

Implication of research and recommendations for future research We have less knowledge of how responsible management practices are emerging in non-western context such as Sub-Saharan Africa and filling up this deficiency in knowledge could significantly contribute to scientific knowledge and theoretical models in the responsible management discipline. Future studies could explore how the stakeholders in such contexts participate in the process.

ABSTRACT. 1. Introduction and Research Context

Evaluating research actors is a critical yet complex task, particularly as the scope of science and technology expands to include diverse contributions such as economic benefits, knowledge creation, social welfare, and policy influence (Stephan, 2012; Bornmann, 2012; Reed, et al., 2021). Traditional evaluation frameworks often emphasize input-output metrics like publications, patents, and citations, which fail to fully capture the multi-dimensional roles and impacts of research actors (Godin, 2007; Lepori, et al., 2008). Furthermore, these traditional frameworks struggle to address intangible factors like relationships and systemic contributions of research organizations (Lepori, et al., 2008; Bornmann, 2012). To address these limitations, this study introduces and develops the concept of positioning of research actors as a new framework for research evaluation. Drawing from the case of Korean Public Research Institutes(PRIs), this paper demonstrates how interdisciplinarity and research collaboration can be used to explore and describe the positioning – roles, functions, and impacts – of research actors within the innovation system. These dimensions are used to examine both tangible and intangible aspects of research actors, offering a more nuanced and holistic understanding of their contributions to the innovation system.

2. Theoretical Framework and Objectives

The positioning of research actors encompasses their roles, functions, and impact within the innovation system, moving beyond simple productivity metrics to emphasize the diverse contributions as well as of research entities (Barré, 2006; Lepori, et al., 2008). Korean PRIs provide an illustrative case study, as they have transitioned from acting as technology importers and disseminators during the country's industrialization period to requiring new roles in a maturing innovation ecosystem alongside private enterprises and universities. This research aims to: • Develop a multi-dimensional framework for assessing the positioning of research actors. • Use interdisciplinarity and research collaboration patterns to explore PRIs’ roles, functions, and impact • Propose actionable insights for designing evaluation frameworks that align with evolving research priorities and policy goals, moving beyond quantitative metrics

3. Methodology

The study used both quantitative and qualitative approach to assess interdisciplinarity and research collaboration as key indicators of positioning: • Interdisciplinarity Analysis: Utilizing the framework of variety, balance, and disparity (Stirling, 2007), this analysis examines how research disciplines are distributed, integrated, and interrelated within PRIs’ research portfolios. While variety and balance were calculated based on National Standard Classification of Science & Technology used by the Ministry of Science, Technology and ICT Korea, disparity is complemented by qualitative evaluations of disciplinary characteristics to account for contextual differences across disciplines (Sugimoto & Weingart, 2014; van Dam, 2019). • Research Collaboration Analysis: Topological network and cluster analyses were used to assess collaboration patterns and the roles of research partners within the network. Additionally, this paper considered varying contributions of partners (e.g., universities, industries) to highlight the different assets each PRI sought through their networks. This combination of complementary methods enables nuanced insights into the positioning of PRIs by shedding light on different aspects of their capabilities, performances, and relationships within the broader science, technology, and innovation(STI) system.

4. Key Findings

The findings highlight distinct roles and functions of Korean PRIs based on their interdisciplinarity and collaboration patterns: • Interdisciplinarity Analysis: PRIs are categorized into roles such as fostering and supporting, specialized, and developing. For instance, PRIs with balanced interdisciplinarity profiles often play supporting roles that nurture diverse research activities across other developing PRIs and SMEs. • Research Collaboration Analysis: Collaboration patterns reveal roles such as bridging institutions that connect diverse scientific fields or specialized actors focusing on applied or experimental sciences. • Alignment and Divergence: Interestingly, PRIs that act as bridging institutions in collaboration often align with fostering and supporting roles in interdisciplinarity. Conversely, specialized PRIs predominantly focus on applied sciences rather than basic research, which contrasts with national strategies emphasizing the development of basic sciences.

5. Implications and Contributions

The results underscore the utility of positioning as a conceptual and evaluative framework. Key insights include: • Policy Alignment: By revealing discrepancies between national research priorities and institutional roles, the framework highlights areas for strategic alignment, such as increasing support for specialism in basic science in PRIs. • Comprehensive Evaluation: Comprehensive Evaluation: Positioning integrates tangible and intangible dimensions, addressing gaps in traditional metrics and offering a holistic view of research actors' systemic contributions. For example, a PRI could play an important role in development of a niche research area, despite a low number of publication or citation due to the nicheness of the research area. • Strategic Guidance: The nuanced insights from positioning indicators can inform policies that optimize the roles of research actors in achieving diverse STI goals, from economic growth to knowledge creation and societal welfare.

6. Conclusion

The concept of positioning of research actors offers a robust framework for addressing the complexities of research evaluation. By integrating interdisciplinarity and collaboration patterns, this approach moves beyond conventional metrics to provide a comprehensive understanding of research actors’ roles, functions, and impacts. Applied to the case of Korean PRIs, the framework demonstrates its potential to uncover multi-dimensional contributions and inform more nuanced and effective research policies. This work not only advances research evaluation methodologies but also supports the broader goal of fostering sustainable and impactful innovation systems.

7. Bibliography

Barré, R.,2006. Towards a European STI Indicators Platform(ES-TIP). Paris, PRIME annual conference. Bornmann, L.,2012. Measuring the societal impact of research. EMBO reports, 13(8), pp.674-676. Godin, B.,2007. Science, accounting and statistics: The input-output framework. Research Policy, Volume36, pp.1388-1403. Lepori, B., Barré, R. & Filliatreau, G.,2008. New perspectives and challenges for the design and production of S&T indicators. Research Evaluation, 17(1), pp.33-44. Reed, M. S. et al.,2021. Evaluating impact from research: A methodological framework. Research Policy, 50(4), pp.104-147. Stephan, P.,2012. How Economics Shapes Science. 1 ed. Boston:Harvard University Press. Stirling, A.,2007. A general framework for analysing diversity in science, technology and society. Journal of the Royal Society Interface, Volume10.

ABSTRACT. Introduction As nations develop economically, environmental concerns have become more prominent. Historically, progress was often equated with development, overlooking environmental impacts. The excessive use of natural and finite resources has contributed to climate change.

To address this, diverse efforts have emerged to mitigate human impact on nature, involving companies, universities, and funding agencies in promoting sustainable research. This study examines the influence of three programs funded by the São Paulo Research Foundation (FAPESP) in Brazil: BIOEN (Bioenergy Research Program), BIOTA (Biodiversity Research Program), and RPGCC (Global Climate Change Research Program). We assess their global contributions by analyzing articles linked to these programs through DOIs provided by the FAPESP virtual library.

BIOEN: Bioenergy Research Program BIOEN fosters research and development in bioenergy, integrating academic and industrial efforts to advance knowledge and applications in Brazil. Its objectives include nurturing academic inquiry, training skilled professionals, and encouraging collaborative research among São Paulo universities, research institutes, and corporations.

BIOTA: Biodiversity Research Program Launched in 1999, BIOTA aims to study, map, and evaluate São Paulo State's biodiversity, contributing to forest conservation policies and sustainable exploitation of plants and animals. With over 1,200 experts, the program focuses on inventorying biodiversity, defining conservation strategies, and exploring economic and sustainable uses.

RPGCC: Global Climate Change Research Program RPGCC advances knowledge in climate change to inform policy development and technological solutions for mitigation and adaptation. The program emphasizes regional and paleo-climatic observations and explores diverse areas such as ecosystem impacts, greenhouse gas emissions, agriculture, human health, and socio-economic responses to climate change.

Preliminary Results Data from the FAPESP virtual library and Dimensions search engine revealed:

BIOEN: Of 130 DOIs, 51 publications were identified, primarily in Agricultural, Veterinary, and Food Sciences, Biological Sciences, and Environmental Sciences. These publications formed seven conceptual clusters with 97 concepts and 1,217 co-occurrence links. The authorship network included 32 researchers, mostly from Brazilian institutions, collaborating with 17 countries across 81 co-authorships. BIOTA: Of 328 DOIs, 128 publications were found, with Biological Sciences, Ecology, and Environmental Sciences as the most represented fields. Five conceptual clusters emerged from 100 concepts and 1,770 co-occurrence links. Brazilian researchers dominated, with collaborations across 37 countries forming 183 co-authorship links. RPGCC: Of 492 DOIs, 113 publications were identified, focusing on Earth Sciences, Atmospheric Sciences, and Environmental Sciences. These formed four clusters with 100 concepts and 1,554 co-occurrence links. The authorship network included 77 researchers and 395 co-authorship links, with most contributions from Brazil, followed by the United States, the United Kingdom, and Germany. The University of São Paulo played a pivotal role in supporting these studies, alongside institutions like INPE (National Institute for Space Research) and São Paulo State University. International collaboration, while present, remains limited, highlighting potential for broader partnerships to address global environmental challenges.

Conclusion Our findings underscore FAPESP’s significant role in advancing environmental research through these programs. The dominance of Brazilian institutions reflects strong domestic collaboration but also points to opportunities for expanding international cooperation to enhance research outcomes and tackle global sustainability issues.

ABSTRACT. Under-representation of women in STEM has been a persistent issue in S&T policies of many countries. With rising attention to the implications of gender diversity for research performance and innovation, this study investigates the moderating effects of gender diversity in linking investment and outcomes of R&D in the South Korean context in an attempt to find out how gender diversity mediates the efficiency of R&D investments for innovative outcomes. Answering this question will not only provide insights for policies for women in STEM but also for innovation studies and policies. Gender diversity in research performance has been extensively studied, though findings remain inconsistent. In the South Korean context, Han et al. (2021) found a negative relationship between gender diversity and both the number of SCI publications and patent applications in government-funded research institutes. Conversely, Kim and Hwang (2022) observed a positive association between gender diversity and intellectual output in the South Korean energy sector. More broadly, gender diversity shows mixed impacts on innovative outcomes. While it significantly influences technological innovation in automobile companies (Hyun, 2023), studies have reported an inverted U-shaped relationship with innovation outcomes (Nakagawa and Schreiber, 2014; Sastre, 2014; Capozza and Divella, 2023). In general, previous studies have focused on the direct effects of gender diversity on research output and innovation outcomes; very few have explored its role of moderating R&D inputs and outputs. This gap calls for a deeper investigation into the interplay of R&D investments, gender diversity, and performance outcomes. Using the data on government-funded research projects from 2017 to 2022 from the National Science and Technology Information Service (NTIS) of South Korea, this study examines how the effects of government funding on various project outputs such as publications, patents, and technology royalties vary by gender diversity of the research project team. A mixed-effects modeling approach utilizing fixed and random effects is employed to address the hierarchical structure of panel data. With the research funding as an independent variable, the fixed effects model tests a moderating variable – gender diversity of the research team against control variables such as the principal investigator’s gender, research team’s educational diversity, institution type, project stage, project field, etc. Random effects are included for projects and years to capture unobserved heterogeneity. The models also apply the time lags of zero to three years between research funding and outcome production, based on previous studies exploring the relationship between research inputs and outputs (Lee and Choi, 2015; Subramanian et al., 2016; Xie et al., 2020). Plus, weighting is applied to single-year projects to address different project periods to prevent underestimation of variation caused by multiple replicated values. Our preliminary findings reveal the critical roles of gender diversity in understanding the returns on R&D investments, providing evidence-based implications for integrating gender diversity into R&D and innovative strategies. Furthermore, the findings highlight that promoting gender diversity is not simply a matter of gender equality but can be a crucial strategy for innovation.

ABSTRACT. The perception that science is becoming increasingly internationalized and interdisciplinary is widespread, yet quantitative evidence at the disciplinary level remains limited. In this study, I leveraged comprehensive digital repositories of scholarly metadata to investigate trends in interdisciplinarity and internationalization across different scientific disciplines. Focusing on well-established, high-impact, peer-reviewed journals in biology, chemistry, economics, medicine, physics, political science, and multidisciplinary sciences, I analyzed the publication records to discern patterns over time. Through mining data from the large-scale, open-source bibliometric dataset OpenAlex and the Clarivate Analytics' Journal Citation Reports for disciplinary classification, I selected journals based on their long publishing history, high annual article output, significant impact factors as of 2022, and emphasis on original research over review articles. This selection included seven multidisciplinary journals and sixteen journals for each of six disciplines across the natural sciences (chemistry and physics), life sciences (biology and medicine), and social sciences (economics and political science). These findings support the perception of a nearly universal trend toward increasing internationalization in both multidisciplinary and discipline-focused journals. However, notable disparities exist among disciplines. Medicine journals are less internationalized compared to other fields and show little increase in internationalization over time. In contrast, physics journals exhibit a segregation between highly internationalized and less internationalized journals. Additionally, while multidisciplinary journals have experienced significant shifts in their disciplinary focus over the past century, disciplinary journals have largely maintained consistent levels of interdisciplinarity. These results highlight the complexity of trends in scientific publishing. While international collaboration is generally on the rise, the extent and rate of this increase vary significantly between disciplines. Understanding these patterns is crucial for policymakers, publishers, and the scientific community as they navigate the evolving landscape of global science.

ABSTRACT. The transformative innovation policy (TIP) frame commits to plurality of interventions, actors, and governance as necessary elements to elicit societal change. However, the interpretive work that policymakers engage in as they operationalize this plurality has been largely overlooked. Through a qualitative study of the U.S. National Nanotechnology Initiative (NNI), this paper explores how national policymakers interpret the plurality commitment in TIP. For over 20 years the NNI has brought together a diversity of actors and invested more than 43 billion USD cumulatively to build a nanotechnology research and development project in the U.S. supported by targeted education and workforce efforts. Drawing on strategic documents from 1999-2023, interviews with policymakers, and participant observation, this analysis identifies five interpretations of plurality in the NNI. When interpreting plurality as multiplicity it emerges as an assemblage of necessary ingredients; interpreted as integration plurality blurs disciplines and sectors; interpreted as transformation it identifies specific groups and values to be included; interpreted as complexity plurality grapples with the challenges of a broad policy landscape; while plurality interpreted as balance aims for an equilibrium of elements within nanotechnology. These findings show how these interpretations may sit uncomfortably with existing national framings and suggests that how policymakers interpret and justify the plurality commitments of TIP shapes the conditions and progress of this policy frame. Bringing to light the political and social implications of plurality in TIP offers a novel way to understand the contingency of such commitments in innovation policy and highlights their implications across emerging technologies.

ABSTRACT. Research questions: The operationalization of mission-oriented innovation policy (MOIP) is a significant area of science, technology, innovation policy, and transformation studies that requires further investigation. This study addresses this crucial gap by conducting a comparative case study on the German Energy Research Program (ERP) and the Energy Earthshots Initiative (EEI) through in-person interviews and document analyses. The ERP has already presented five distinct missions to foster the transformation of the German energy sector towards a net-zero goal in 2045. The Biden-Harris administration assembled eight “shots” to reach the same goal for the United States. The DoE, its ARPA-E, and other Departments and Agencies execute the EEI by funding research projects nationwide, from basic research to more mature technology. With a conceptual background in evolutionary economics (Nelson und Winter 1985) and experience in transformation studies a project at the professorship has already analyzed the governance of the German case. However, the focus on the EEI’s governance presents a unique opportunity for further investigation and a perfect comparison case. As both energy innovation research programs can be perceived as mission-oriented, this research design offers a novel and unique opportunity for an inductive approach to how mission orientation is or could be operationalized. For this purpose, professionals in the EEI realm will be interviewed, and all relevant documents regarding the EEI will be coded jointly with the interviews' transcripts. Primarily, the case of the EEI could provide valuable insights into how to push forward energy transformation, a prospect that we eagerly anticipate.

• Which specific processes are pivotal in the governance framework for implementing the "Energy Earthshots Initiative"? • In what manner is the governance framework structured within the "Energy Earthshots Initiative," encompassing its organizational components, decision-making processes, and mechanisms for stakeholder engagement and coordination?

Methodology: The suitable research design for this purpose is a qualitative case study grounded on in-depth interviews, aligning with the methodological framework by Yin (2018) and a document analysis (Miles et al. 2014; Saldaña 2013).

Literature review: As stated above, there is a conceptual gap in the literature on MOIP. There are many studies on the U.S. National Innovation systems. Moreover, there are studies on technologies included in the EEI with a focus on including US cases (e.g.,Andreasen und Sovacool 2015; Wachs et al. 2023; Woodall et al. 2022). What is more, studies are providing a nuanced overview of the innovation policy landscape of the United States (Bonvillian 2022; P. Shapira et al. 2010; Liu 2022) and some, which shed light on the energy innovation policy in particular (Stevens et al. 2023; Popp 2020). Taking one step further in this research, Arnulf Grübler et al. (2012) investigate the governance of U.S. policy. Concerning comparing Germany and the United States, the work by Hommes et al. (2010) is the most recent update on the general NIS. For the ARPA-E, there are some excellent studies, even based on interviews in recent years (Azoulay et al. 2019; Bonvillian und van Atta 2011; Carleton und Cockayne 2023; Haley 2017; National Academies of Sciences, Engineering, and Medicine 2017; Watson 2022). In sum, some studies examine most of this research's pillars. However, more literature is needed on the governance of energy innovation policy in light of the EEI and more landscape political changes like the IRA. Moreover, the recent changes in the political landscape (Geels 2020) add the issue of the robustness (Sørensen und Ansell 2023; Sørensen und Torfing 2024; Trondal et al. 2022; Ansell 2024) of the EEI and its executing actors to the lenses analyzing the governance of the EEI.

Findings: Since the interviews have just started the findings based on the interviews are preliminary. The document analysis reveals some interesting insights already. The shots have been announced over the course of three years and encompass mostly technology-driven cost or efficiency targets which feed in the grand societal challenges of decarbonization. A minimum of two offices, under the umbrella of the undersecretary for science and innovation, are spearheaded on a shot. Since all shots are connected to existing DoE programs, there is no additional funding in place but additional goal formulation and alignment of the efforts. The newly founded Energy Earthshot Research Centers (EERCs) are led by a DOE National Laboratory and display the cross-office character of the EEI. Some shots, for example, the floating off-shore shot, include other departments. Thus, coordination, reflexivity in terms of feedback, and the continuous alignment of the goals to the grand challenge are core tasks and encounters for the governance framework of the EEI. The inaugural summit organized by the leading office expresses that those principles of MOIP are constantly accounted for. Those events function as a tool to spur first informal and later on formal coordination mechanisms in the community working on a specific shot and provide the opportunity for feedback along the vertical and horizontal structure of actors. Additionally, regular shot-specific teams across the DoE offices convene in meetings to align their activities and discuss new funding opportunities or requests for information, etc., to foster the horizontal coordination of the policy mix. Other actors within the DoE play smaller but potentially crucial roles in executing the EEI. For example, ARPA-E supports the technologies and potential projects that could feed into the shots. It has a direct connection to practitioners working on the “next big thing.” The loan programs could provide crucial funding to assist technologies or projects closer to market entry because it is not bound to TRL 1-3. The reaction of EEI’s actor eco-system to the political landscape shift cannot be evaluated at this point to a full scale. Looking at the threats the ARPA-E encountered during the last Trump presidency and how it overcame them, it is not certain that the outlook is indicating EEI is going to cease to exist. Likely, there will be budget cuts and changes in the personnel (cuts) and strategies of robustness, for example rebranding of some parts of the program or highlighting the pillars suitable for the new political agenda.

ABSTRACT. Extracting lunar resources involves navigating the legal complexities of the Outer Space Treaty (OST) while balancing the interests of civil, commercial, and military stakeholders. This is particularly challenging in resource-rich areas of the Moon, where overlapping U.S. and Chinese mission plans could lead to geopolitical tension. Implementing robust governance mechanisms is crucial to upholding non-appropriation principles, mitigating conflicts, and promoting operational harmony. While the Artemis Accords propose “safety zones” to prevent interference, their suitability for large-scale, multinational operations remains uncertain. These zones must balance exclusivity with equitable access to ensure long-term mission efficiency.

The Artemis Accords used the protection principle of safety zones to avoid any interference in lunar activities. However, such zones may have to be expanded and fine-tuned to accommodate commercially viable mining and long-term operational complexities. In my previous work, ISRU (In-Situ Resource Utilization) investigated the economic effects, environmental concerns, and safety considerations of lunar mining compared to terrestrial supply. Using a single-nation model, I identified efficient, low-impact extraction zones and demonstrated gradual resource depletion. This research highlighted the potential for ISRU to stabilize costs despite high initial investments.

Challenges such as overlapping claims and proximity-driven conflicts emerge as lunar exploration shifts to include multiple actors. Terrestrial frameworks, such as the UN Convention on the Law of the Sea, provide valuable insights for managing shared resources. Non-governmental entities such as the Open Lunar Foundation have also called meetings on improving safety zone proposals, which present possible channels for multi-lateral regulation. These frameworks imply that integrated working relationships are required to set limits and manage challenges in a collectively used lunar space.

This research addresses a key gap by integrating multi-actor dynamics into ISRU simulations, providing policymakers with data to evaluate safety zones and shared resource sites. My work will focus on modeling the efficiency and scalability of these zones under various geopolitical and operational scenarios. The simulation will examine critical metrics, such as spatial overlap and proximity impacts, to provide actionable insights for designing equitable and effective resource-sharing frameworks.

My prior ISRU simulation will be the foundation for a multi-actor model incorporating variables such as resource demand, boundary conflicts, and operational scaling. The first simulations will involve two agents, representing a basic geopolitical landscape where the agents’ missions intersect. Data from the Lunar Reconnaissance Orbiter and PSR mapping missions will inform these models' spatial and environmental dynamics. The single-nation model will serve as a comparative baseline, allowing for a clear analysis of the added complexities introduced by multi-actor interactions.

As lunar exploration and the use of space resources increase, creating proper governance in this area becomes critical. This research thus benefits sustainable and equitable space policy by simulating how safety zones and shared-resource sites can promote cooperation and reduce conflicts. This work provides actionable data to civil, commercial, and military stakeholders, forming a basis for agreements that promote resource sustainability and operational harmony. It emphasizes the need for flexible governance to ensure fair resource access and reduce tensions in shared lunar environments.

ABSTRACT. The growing fleet of satellites in Earth’s orbit are critical to daily life, yet satellites remain the only complex infrastructure without regular maintenance, repair, or upgrade capabilities. With this motivation, the commercial, military, and academic sectors are developing technologies for in-space servicing, assembly, and manufacturing (ISAM). A paradigm shift is imminent: routine service to extend the operational lifetime of space-based assets is necessary to ensure a reliable, resilient, and sustainable space domain.

In general, ISAM refers to the suite of technological capabilities used in Earth’s orbit, cislunar space, deep space, or on the surface of space objects and celestial bodies. The “servicing” aspect of ISAM is characterized by activities that require a close maneuver and operation in the near vicinity of a customer spacecraft. Potential services include in-space inspection, refueling, station keeping, life extension, repair, and mechanism deployment. The successful coexistence of customers and the ISAM infrastructure elements is enabled precise communication and coordination is required without interfering with nominal operations. This requires careful radiofrequency spectrum management.

ISAM infrastructure design is nascent. ISAM infrastructures will require many different communication links to enable operations. The FCC is currently laying the groundwork to establish a regulatory framework for the radiofrequency spectrum as it relates to ISAM operations. In FCC 24-21, issued in February 2024, the FCC has tentatively decided to evaluate ISAM operators’ spectrum needs on a case-by-case basis. As servicing becomes a routine aspect of operations in space, this approach may not be sufficient to support the industry. Understanding the communication requirements of potential infrastructures is an active point of discussion in the space policy field. This work explores historical and proposed ISAM missions in an effort to quantify and qualify the communication requirements for different ISAM infrastructures. Recommendations for policy and decision makers are provided as steps towards establishing a concrete spectrum management framework.

ABSTRACT. On October 19, 2024, Intelsat reported that an anomaly occurred on the Intelsat 33e satellite, resulting in a loss of power. Later on, on October 21, 2024, Intelsat confirmed that this anomaly had resulted in the spacecraft breaking up in geosynchronous equatorial orbit (GEO), resulting in a total loss.[1] While troublesome on its own, as space debris in GEO threatens other high-value satellites in that regime, this incident is particularly noteworthy because it follows the breakup of Intelsat 29e in 2019, both of which used the same Boeing platform.[2] While hazardous satellite breakups have occurred in the past, as of 2021 throughout all orbital regimes around 560 incidents have occurred, with only 67.3% of these events being nondeliberate making these events rare.[3] However, for such an incident to happen to the same satellite operator with satellites that utilized the same bus and engine warrants further investigation.[2] While it is certainly possible that it could be just a coincidence, caused by micrometeorite impact or a short circuit, in light of recent controversies surrounding Boeing[4] there are concerns about the seven other GEO satellites that share the BSS-702MP platform.[2] However, it is worth noting that the BSS-702MP is only one variant within the extensive BSS-702 family[2], and even that fails to account for the myriads of other spacecraft suppliers that have their own families of satellite buses. This raises the question: How many satellites in GEO share the same buses? What about sharing the same components? Is part and platform sharing commonplace in GEO? Understanding these patterns contributes to critical risk analyses in the domain. Utilizing a variety of publicly available spacecraft databases and system manufacturing documentation, this report will tabulate their information and map the interconnectivity of parts and suppliers used for satellites throughout the current space environment.

While there has been past literature looking at platform sharing with regard to NASA’s and SpaceX’s launch architecture alongside work noting the flood of mass-produced satellites into space, little to no work has been done to characterize this growth over time within the space industry.[5] A cursory analysis of over 30,000 cataloged satellites and their engine manufacturers using data provided by Seradata, a commercial space data aggregator, shows that major GEO players such as Intelsat and SES have relied on engine suppliers who have also worked on at least another 2,000 satellites. Additionally, while Nammo, the engine supplier for Intelsat 29e and 33e,[2] has supported five separate systems, other suppliers used by SES and Intelsat, such as OKB Fakel and Aerojet Rocketdyne, support as many as 851 and 574 satellites, respectively. How this trend continues for the individual engines and for other critical components, such as batteries or propellant tanks, requires further study. While Intelsat has a track record of transparency regarding anomalies on its spacecraft, other entities may attempt to obfuscate the situation presenting in which any number of satellites could be ticking time bombs, set to go off with little to no warning. This presents not only a critical obstacle for space sustainability efforts but also a credible national security risk to GEO assets critical for early missile warning, nuclear command and control, and signal intelligence. By characterizing and mapping out the current GEO environment, it can pave the way for solutions to these important problems.

1. Stevens, Maureen. “Intelsat Reports IS-33E Satellite Loss.” Intelsat, October 31, 2024. https://www.intelsat.com/newsroom/intelsat-reports-is-33e-satellite-loss/.; Rainbow, Jason. “Intelsat 33e Breaks up in Geostationary Orbit.” SpaceNews, October 21, 2024. https://spacenews.com/intelsat-33e-loses-power-in-geostationary-orbit/. 2. Krebs, Gunter D. “Hughes / Boeing: HS-702 / BSS-702, HS-GEM / BSS-GEM (Geomobile)”. Gunter's Space Page. Retrieved November 21, 2024, from https://space.skyrocket.de/doc_sat/hs-702.htm; Roberts, Thomas G. “The Commercial Space Industry (Part I),” Lecture, Georgia Institute of Technology, Habersham Building, Atlanta, October 24, 2024. 3. Ren, Si-yuan, Zi-zheng Gong, Qiang Wu, Guang-ming Song, Qing-ming Zhang, Pin-liang Zhang, Chuan Chen, and Yan Cao. “Satellite Breakup Behaviors and Model under the Hypervelocity Impact and Explosion: A Review.” Defence Technology 27 (September 2023): 284–307. https://doi.org/10.1016/j.dt.2022.08.004. 4. Chappell, Bill. “How Bad Is Boeing’s 2024 so Far? Here’s a Timeline.” NPR, March 20, 2024. https://www.npr.org/2024/03/20/1239132703/boeing-timeline-737-max-9-controversy-door-plug. 5. Ansar, Atif, and Bent Flyvbjerg. “A Platform Approach to Space Exploration.” Harvard Business Review, September 28, 2023. https://hbr.org/2022/11/a-platform-approach-to-space-exploration.; Roeder, Tom. 2019. Space symposium 2019: Space race going hypersonic with mass-produced satellites, falling costs. TCA Regional News, Apr 14, 2019. https://www.proquest.com/wire-feeds/space-symposium-2019-race-going-hypersonic-with/docview/2209100241/se-2 (accessed November 21, 2024).]

ABSTRACT. In the Cislunar environment, five Lagrange points uniquely offer near-stationary orbits in the Earth-Moon frame, making them strategically valuable for scientific, commercial, and logistical space activities. Among these, the Earth-Moon Lagrange points host quasi-stable orbits such as Halo orbits and Lyapunov orbits, which present a heightened localized potential for collision. The growing interest in utilizing these orbital regions, coupled with the absence of formalized communication and coordination mechanisms among spacefaring nations, raises critical concerns about collision risks. As the density of assets in the lunar sphere increases, this issue will become a pressing matter, demanding preemptive legislative infrastructure to mitigate risks. This paper explores and documents guidelines for international communication and collaboration in proximity to the Earth-Moon Lagrange points. It identifies how current intergovernmental treaties and regulatory frameworks, designed primarily for Earth-centric space activities, inadequately address the complexities of this evolving domain. Specifically, it examines the limitations of existing infrastructure in managing operations near the Lagrange points, highlighting the absence of regulations suited to this environment. Through a high-fidelity simulation environment, this study quantifies the collision risk associated with future missions. The simulations is based on real provisional mission profiles. Analysis suggest that satellites placed in more stable orbits, often a preference due to commercial profitability requirements, face significantly higher risks of collision. These findings underscore the importance of coordination amongst operators to reduce risk while promoting sustainable growth of lunar exploration activities. Building on this quantitative analysis, the paper proposes amendments to existing international agreements, including the United Nations’ Registration Convention and the ITU’s regulatory framework. Recommendations are made to incorporate specific metrics for this unique orbital category into the Registration Convention, ensuring accurate and detailed reporting of satellite orbits. Additionally, the paper advocates for an expansion of the ITU's framework to include operational guidelines tailored to Lagrange point applications, such as quantifiable thresholds for permissible orbital proximity and collision avoidance. The analysis outlines suggestions for threshold design in this unique orbital environment to coordinate orbiting with close proximity. In conclusion, while the current level of activity near the Earth-Moon Lagrange points does not yet pose immediate collision risks, historical precedents in space operations suggest that this could change rapidly. Establishing robust guidelines for communication and coordination is essential to avoid future crises. As human and robotic presence in the Cislunar sphere expands, ensuring the safe coordination of assets near these points will be critical. This paper contributes to the ongoing discourse by providing actionable recommendations to advance international regulatory frameworks and promote long-term sustainability in the Cislunar environment.

ABSTRACT. Artificial Intelligence (AI) is a rapidly advancing technological field that promises to have a direct and positive effect on productivity (Brynjolfsson et al., 2019). AI deployment can drive firm productivity through several channels, by reducing costs, creating new products or services, optimizing resource allocation, and improving precision in decision-making (Acemoglu & Restrepo, 2019; Aghion et al., 2019)). Although studies increasingly focus on understanding the productivity effects of AI, current research predominantly relies on indirect estimates of AI adoption, using firm-level surveys (Rammer et al., 2022), vacancy data (Acemoglu et al., 2022), or patent data (Damioli et al., 2020). While more direct worker-level insights have been gleaned from experimental research (Li & Raymond, 2023; Noy & Zhang, 2023), due to lack of granular firm-level studies, little is known about the productivity effects of AI integration into firms’ production processes. As a result, capturing the underlying mechanisms and estimating the impact of AI adoption on workers remains an empirical challenge.

To provide evidence on this, we exploit a staggered rollout of an AI into a specific production task at a multinational manufacturing firm. The task we study is a quality assurance (QA) step in the production of circuit boards destined for the automotive industry. QA is particularly critical for the firm as many of the circuit boards can enter safety critical functions such as air bags. The production line is highly automated. Robots place components and solder onto circuit boards, ovens then melt the solder to attach components. A high resolution camera captures images of the board, these are processed by a digital, rules-based image processor (AOI) to classify images as pass or fail. Due to high quality standards, the AOI often fails to reach the threshold for classification. Therefore, images are then classified by human verification operators. To enhance QA efficiency and performance, the firm introduced a supervised learning algorithm to partially automate this optical classification task across its plants. We use data provided to us by the firm to causally identify and estimate the direct effect of the AI roll-out on labour demand, task productivity, and human decision making.

Our unique dataset contains information on the classification of approximately 188 million images made by operators and/or AI, spanning an 18-month period (2022-24). The data covers 2 phases of AI implementation: Advisory AI (Phase 1) and Autonomous AI (Phase 2). In the Advisory phase, if the operator decision deviates from AI’s, operators are alerted via a pop-up message and asked to revise their decision. In the Autonomous phase, AI takes over the decision-making and autonomously classifies passing images. Images that are classified as fail, or where the AI is unable to reach the threshold, are then blind reviewed by operators. We rely on the variation introduced by this staggered deployment of AI phases (from Advisory to Autonomous) across plants and analyze its impact on workload reduction, productivity, human decision making, and task complexity.

We expect to show that the AI reduces the labour demand for human operators, as the direct result of the automation of the classification task. The anticipated effects for overall productivity are more ambiguous: whilst we expect the AI to increase the speed of the QA process; other effects, such as changes in error detection, could dampen efficiency savings. Finally, we aim to demonstrate a small, but significant, influence of the AI on human decision making when it is in Advisory Phase.

This study makes several contributions to understanding AI’s role in reshaping work and productivity in manufacturing. First, by focusing on task-level data, we offer a granular view of AI’s impact on human work within a specific operational context, moving beyond aggregate productivity metrics. Second, our findings on workload and decision dynamics have broader implications for labor substitution and augmentation, illustrating how AI can either reduce or reconfigure human tasks, whilst also furthering a process of automation in manufacturing. For practitioners, our analysis highlights best practices for implementing AI in critical, high-stakes quality control processes. The staged approach (from Advisory to Autonomous AI) provides valuable insights into gradual AI deployment strategies that balance operational efficiency with the need for human oversight.  

References    

Acemoglu, D., & Restrepo, P. (2019). Automation and new tasks: How technology displaces and reinstates labor. In Journal of Economic Perspectives (Vol. 33, Issue 2, pp. 3–30). American Economic Association. https://doi.org/10.1257/jep.33.2.3

Acemoglu, D., Autor, D., Hazell, J., & Restrepo, P. (2022). Artificial Intelligence and Jobs: Evidence from Online Vacancies. Journal of Labor Economics, 40(S1), S293–S340. https://doi.org/10.1086/718327/SUPPL_FILE/20462DATA.ZIP 

Aghion, P., Antonin, C., & Bunel, S. (2019). Artificial intelligence, growth and employment: The role of policy. Economie et Statistique, 2019(510–512), 149–164. https://doi.org/10.24187/ecostat.2019.510t.1994

Brynjolfsson, E., Rock, D., & Syverson, C. (2019). Artificial Intelligence and the Modern Productivity Paradox. In The Economics of Artificial Intelligence (pp. 23–60). https://doi.org/10.7208/chicago/9780226613475.003.0001

Damioli, G., Vincent, ·, Roy, V., & Daniel Vertesy, ·. (2020). The impact of artificial intelligence on labor productivity. Eurasian Business Review, 11, 1–25. https://doi.org/10.1007/s40821-020-00172-8

Li, E. B. D., & Raymond, L. R. (2023). GENERATIVE AI AT WORK. NBER Working Paper, 82–95.

Noy, S., & Zhang, W. (2023). Experimental evidence on the productivity effects of generative artificial intelligence. Science, 381(6654), 187–192. https://doi.org/10.1126/science.adh2586

Rammer, C., Fernández, G. P., & Czarnitzki, D. (2022). Artificial intelligence and industrial innovation: Evidence from German firm-level data. Research Policy, 51(7), 104555. https://doi.org/10.1016/J.RESPOL.2022.104555

ABSTRACT. This research addresses the critical issue of design and development of products like production machines for SMEs and rural communities as an essential capability to foster local technology development within the innovation systems of Latin America. What is studied and discussed in this research is the pre-requisites for effective application of design and development methods, like design thinking for prototyping and industrial production machinery, in collaborative spaces of universities and small and medium enterprises (SMEs) in the Bolivian context. For this purpose, the theoretical and empirical perspectives of central characteristics and critical success factors for design thinking implementation for prototyping and product development for SMEs clusters are studied and evaluated. The SME cluster initiatives are facilitated by a public university, which follows the mission of a developmental university through the democratization of knowledge, with one of its principal activities being the design and development of products like industrial production machines. The adoption of design thinking approaches and methods has been introduced as a new tool in the supporting activities between university-industry, to strengthen the SMEs’ and rural communities’ capabilities to design, prototype and develop new industrial production machines and new agricultural production methods. This research-based framework may facilitate the SME managers’ understanding of how it works and how it can be applied successfully, which is particularly valuable for resource-constrained SMEs. The framework shows central characteristics of design thinking implementation like dimensions of critical factors, strategies, tools, and phases. Based on the identification of the critical factors some strategies emerged to improve the development of prototypes and machines like the use of visualization tools, such as customer journey maps within SMEs cluster initiatives context. This tool inspires and promotes communication with users and stakeholders, to get a deeper understanding of user needs. This facilitates the achievement of more satisfactory results of feasible, viable and sustainable machine projects that are appropriate to the capabilities of users/clients. In resume, this research elucidates some issues on how facilitate the implementation of design thinking for prototyping and product development. It further explores how this approach can contribute to addressing problems within the context of university-led cluster initiatives involving SMEs and farmers with limited resources. This evidence underscores the broad applicability of design thinking approach and highlights the extensive potential for further research into its implementation within this specific context.

ABSTRACT. International scientific collaboration defines the fourth age of research, with policy incentives often cited as key motivations for researchers to engage in cross-border collaboration and exchange. However, empirical evidence outside the Western context remains scarce, and the heterogeneity within international collaboration is often overlooked. To address these empirical and conceptual gaps, this study examines the impact of Taiwan’s New Southbound Policy (NSP) on its scientific collaboration with eight designated priority countries, including Indonesia, Malaysia, the Philippines, Thailand, Vietnam, Singapore, India, and Australia, over the period 2011–2021. Drawing on bibliographic data from the Web of Science (WoS) Extended API, we analyzed 28,465 co-authored articles. Funding status is determined based on funding acknowledgments and co-authorship types are assigned according to the country affiliations of first and last authors: TWN-led, NSP-led, Equal, Minimal, and Co-affiliated, based on affiliations.

Our initial observations reveal that 75% of the TWN–NSP co-publications received funding, with 52% supported by Taiwan. Overall, 40% of all co-publications were funded by Taiwan’s grant agencies. TWN-led collaboration had the highest share of funding from Taiwan, with at least half receiving support. In contrast, fewer than 25% of NSP-led collaboration received funding from Taiwan, despite having a comparable funding rate of 75%. Minimal collaboration, where neither Taiwan nor NSP priority authors were first or last authors, experienced a sharp decline in funding from Taiwan, dropping from 65% in 2012 to just over 25% in 2021, despite maintaining the highest overall funding rates. Equal and co-affiliated publications had the lowest funding rates, yet 40–50% still received Taiwan’s support. These preliminary results suggest that the centrality of authors in research, signaled through authorship order, may be closely tied to funding.

Results from the piecewise regressions largely confirm our observations. While the rate of TWN-led co-publications funded by Taiwan barely changes, the number of such publications shows a significant increase, averaging 62 papers per year after 2016. All co-authorship types, except Minimal, exhibit significantly lower numbers of co-publications. After 2016, however, the decline is more pronounced for Minimal, NSP-led, and Equal, each with 84, 49, 38 papers less annually. Regarding funding rates, all co-authorship types, including Minimal, are less likely to be funded by Taiwan compared to TWN-led publications. In particular, the funding rate for Minimal and Co-affiliated publications decline at 3% annually before 2016.

Our present analysis provides indirect evidence of the positive impact of New Southbound Policy (NSP) on scientific collaboration between Taiwan and NSP priority countries from a funding perspective, as reflected in the significant boost in TWN-led co-publications after 2016. For future work, we plan to leverage sentence embedding techniques and Research Organization Registry (ROR) data to better identify the countries associated with grant agencies in our dataset. Additionally, we aim to investigate the internal variations across countries. By doing so, we anticipate gaining a more granular understanding of the funding mechanisms driving the NSP initiative in science and technology cooperation.

ABSTRACT. Spaceport Camden, a proposed commercial launch site on the Georgia coast was envisioned in the 1960s as the launch hub for NASA’s Apollo program. After losing out to Cape Canaveral, Florida, the site saw little space-related activity until 2012 when Camden County began pursuing its development as a commercial spaceport. With plans for launches by 2020, the project gained momentum, leading to FAA authorization for vertical launches in 2021. However, persistent challenges, including environmental concerns from nearby residents, led the Georgia General Assembly to dissolve the project in 2024 via H.B. 1489. Spaceport Camden offers a detailed case study of how spaceport development collides with socio-political, economic, and environmental realities. While the project aimed to capitalize on the rapid growth of the private space industry, which has seen annual launch rates increase significantly in the past decade, it ultimately services as a cautionary tale about the barriers that can derail even federally approved initiatives. The most recent notable literature discussing the site is from 2019, prior to FAA authorization and the dissolution of the Camden County Spaceport Authority. This undergraduate research provides an updated chronological “biography” of the spaceport, from its early conceptualization to eventual termination. It further describes the socio-political, economic, and environmental factors which made it unfeasible in Georgia. The poster presentation will include a stakeholder analysis of Spaceport Camden, a 1960-2024 site timeline, and a list of takeaways for commercial actors considering spaceport development. Qualitative data sources include state legislative committee transcripts, state legislation, state and federal policy documents (i.e., white papers, fiscal notes, licensing documentation), think tank literature, media publications, and interviews with state policymakers involved in H.B. 1489. Qualitative data is analyzed using thematic and axial coding to identify categories of factors that led the State of Georgia to reject the spaceport. Quantitative data includes economic impact calculations from state and local government reports and statistics pulled from the literature to contextualize takeaways. This work contributes to the evolving discourse on commercial space launch by describing a spaceport case rarely discussed in literature and extracting meaningful takeaways for private actors.

ABSTRACT. Introduction and Context Renewable energy technologies have emerged as a cornerstone of global efforts to address interconnected challenges of energy access, agricultural sustainability, and inclusive development. In Nepal, where smallholder farmers (SHFs) form most of the agricultural workforce, reliable irrigation is a persistent challenge. Limited water access perpetuates cycles of poverty and food insecurity, constraining SHFs’ ability to maximize yields, adopt high-value crops, and build resilient livelihoods. Solar-powered irrigation systems (SPIs) offer a transformative solution by providing an environmentally sustainable and economically viable alternative to traditional diesel- or electricity-powered pumps. Recognizing their potential, Nepal has implemented a renewable energy policy to subsidize SPI adoption among SHFs. This policy aims to enhance access to irrigation technology, promote sustainable agricultural practices, and improve energy efficiency. However, the adoption of SPIs depends on complex socio-economic and cultural factors, including farmers’ decision-making processes, values, and self-perception within their communities. Despite the global promotion of SPIs, their impact on SHFs in resource-scarce settings remains under-explored. This study examines the interplay between Nepal’s renewable energy policy and SHFs’ lived experiences, investigating how SPIs influence farmers’ agency, identity, and values. It provides insights into how renewable energy policies can advance innovation, inclusivity, and sustainability. Research Objectives This study explores the decision-making processes of SHFs in adopting SPIs, focusing on how their agency, identity, and values interact with policy frameworks and subsidy mechanisms. The research addresses the following questions: 1. How do SHFs exercise agency in adopting SPIs? 2. How does SPI adoption influence SHFs’ identity, particularly their roles as farmers and community members? 3. What values do SHFs associate with SPIs, and how do these values guide adoption and utilization? The study situates SPIs as a key technological intervention supported by Nepal’s renewable energy policy, assessing their broader implications for SHF productivity and livelihoods. Methodology The research employs an interpretive methodology to capture the subjective and socially constructed realities of SHFs. Data collection involved 91 semi-structured interviews conducted across six districts of Nepal, selected to represent diverse agricultural contexts and SPI adoption rates. The interviews, carried out in five phases, provided iterative insights into farmers’ evolving perspectives. Focus group discussions supplemented the interviews, offering collaborative forums where farmers reflected on SPI adoption and its socio-economic and environmental dimensions. Thematic analysis was used to identify shared patterns, narratives, and contextual differences. This mixed-methods approach ensured a comprehensive understanding of how SHFs’ internal decision-making factors (agency, identity, and values) interact with external influences, such as policy frameworks, subsidies, and local infrastructure. Key Findings SHFs’ Agency in Adopting SPIs While 90% of respondents relied on external assistance to navigate subsidy applications, this did not diminish their active engagement in the decision-making process. About 65% first learned about SPIs through informal networks, such as neighbors or community members who had already installed the system. Observational learning played a key role, with 40% of respondents visiting nearby farms to see SPIs in action before deciding to adopt. Formal institutions, such as microfinance organizations and ward offices, informed 25% of respondents. Microfinance institutions were noted by 18% for providing detailed guidance on SPI benefits, including 20-year warranties for solar panels and insurance coverage. For 70% of farmers, low operational costs compared to diesel or electric pumps were decisive in their adoption. Influence of SPI Adoption on SHFs’ Identity Adopting SPIs reshaped SHFs’ self-perception, fostering a sense of empowerment and resilience. Sixty percent of respondents expressed pride in their ability to irrigate independently, even during power outages or water scarcity. For 45%, SPI adoption symbolized progress, enabling a transition from traditional methods to modern, sustainable practices. Around 25% of farmers became local advocates for SPIs, promoting the technology among peers. This peer-driven advocacy reinforced their identity as progressive and forward-thinking contributors to agricultural innovation, highlighting the social ripple effects of SPI adoption. Values Associated with SPIs SPIs were strongly associated with economic efficiency, sustainability, and reliability. Eighty-five percent of respondents cited reduced operational costs as a primary motivator, particularly in areas with high diesel or electricity expenses. Seventy percent valued SPIs for providing a consistent water supply during dry seasons, addressing critical resource shortages. Environmental sustainability also resonated with farmers, with 30% emphasizing SPIs as a solution to water management challenges, reducing reliance on fossil fuels. Additionally, 50% highlighted SPIs’ role in increasing agricultural productivity, with some reporting a 20-30% boost in yields due to reliable irrigation. Farmers noted that this technology enabled year-round farming and supported high-value crops like vegetables, ultimately enhancing household incomes. Factors Influencing the Decision to Adopt SPIs 1. Financial Accessibility: Nearly 80% of respondents cited subsidies as crucial to adoption, reducing installation costs. Post-installation, 85% highlighted the prospect of minimal operational expenses as a key motivator. 2. Reliable Irrigation: Seventy percent valued SPIs’ reliability during electricity outages, particularly for vegetable farming and livestock management. 3. Observational Validation: Forty percent of respondents reported being influenced by seeing SPIs operating successfully in their communities. 4. Environmental Fit: Farmers in regions with declining water resources (30%) appreciated SPIs for maintaining irrigation during dry seasons, ensuring stable production. Broader Implications This study highlights the critical role of renewable energy policy in fostering sustainable agricultural innovation. Nepal’s subsidy program has improved SPI accessibility, empowering SHFs to enhance productivity and resilience. However, findings underscore the need for context-sensitive policy design. For example, while SPIs are effective in areas with sufficient sunlight and water sources, their immobility poses challenges for farmers with fragmented landholdings. Globally, the research provides valuable lessons for policymakers and development practitioners. It demonstrates that renewable energy policies must address social and cultural dimensions alongside technical and economic factors. By aligning with farmers’ agency, identity, and values, such policies can foster inclusivity and ensure equitable benefits for marginalized populations. At the intersection of innovation, sustainability, and inclusivity, this research contributes to the discourse on how science and technology policies can address global and local challenges. Examining SPIs as a case study, it offers actionable insights for designing policies that support smallholder farmers, promote renewable energy, and advance sustainable agricultural development.

ABSTRACT. The ability to communicate from the Earth to satellites and back enables vital technologies such as GPS, climate monitoring, satellite imagery, and broadband internet. However, the radio frequency (RF) spectrum is a limited resource. Two nearby satellites can not communicate on the same frequency without interference from each other’s communications. As satellite constellations and space systems from commercial, civil, and military actors have rapidly increased over the past 10 years, the strain on this limited resource has intensified. The International Telecommunication Union Radiocommunication Sector (ITU-R), a sector of the specialized global agency of the UN that manages the radio-frequency spectrum and satellite orbit resources, establishes regulations for what constitutes as harmful interference. However, this information is scattered throughout over 2300 pages of their Radio Regulations and current literature does not depict changes to ITU-R regulations in a digestible manner, often referencing changes to numerical bands or interference limits that mean little to those without experience in satellite communication policy or technology. This study seeks to provide an enhanced understanding of the evolution of the ITU’s regulations surrounding “harmful interference” through an investigation of actions such as a visualization of the development of ITU allocation agreements within a popular frequency range and the progression of equivalent power flux density (EPFD) limits. This evolution will be contextualized within the development of key technical advancements in space communications technology utilized by various actors in the space industry. To accomplish this, the author will utilize the ITU Space Explorer, a platform that provides access to the Space Networks Systems Database of the Radiocommunication Bureau of the ITU, to construct a spectrum diagram displaying the evolution of space service frequency allocations over time within a small frequency range. Additionally, the 2027 World Radiocommunication Conference (WRC-27) agenda will be analyzed to discuss key agenda items pertinent to the development of modern interference regulations. Relevant current discussions surrounding harmful interference in satellite communications, such as SpaceX’s petition to revise current EPFD limits, will be analyzed to provide a clear description of the present debate surrounding these regulations. This work will facilitate the ability for those engaged in the space enterprise, but not actively involved in spectrum regulation, to understand the impact of key spectrum policy decisions surrounding harmful interference. Additionally, it will allow those familiar with terrestrial radio communication policy but unfamiliar with its space counterpart to communicate effectively between parties, a skill that will be vital to support the developing relations between satellite and terrestrial mobile network operators. An advanced understanding of the evolution of spectrum regulation surrounding harmful interference for both policy makers and technical contributors will promote the development of policy and technology that effectively utilizes and distributes the limited RF spectrum, encouraging a healthy densification as more actors get involved. This will enhance international coordination, sustainability, and security through adequate and equitable access to the RF spectrum for new space systems, resolution of conflicts due to lesser inter-orbit and terrestrial-satellite interference, and increased reliability for vital security assets in space. 

"Other" Role: Undergraduate

ABSTRACT. As the true scale of the proliferation of resident space objects (RSOs) in the last decade becomes clearer, and commercial space operators continue to announce plans for hundred- and thousand-satellite constellations to be launched in the next decade, concern follows that the low-Earth orbit (LEO) environment will become significantly more congested. In 2024, the European Space Agency (ESA) reported that nearly 40,000 RSOs are tracked and maintained in catalogues by current space surveillance networks. The same report estimated that many more space objects remain untracked: more than 40,000 debris objects greater than 10 cm, more than one million between 1 cm and 10 cm, and 130 million between 1 mm and 1 cm [1]. In addition, many companies have declared their plans to grow the RSO population by an order of magnitude, with estimates ranging between 50,000 and 100,000 new objects from just over half a dozen operators [2], leading to concern as to the permanent impacts of space industry growth on long-term LEO sustainability. A key component of mitigating space environment debris generation and maintaining navigability is post-mission disposal (PMD), which refers to the practice of actively (or passively) removing satellites from densely populated orbital regimes at the end of their operational lifetimes. However, PMD policies may not be consistent or rooted in an understanding of the rapidly evolving space environment – internationally agreed-upon guidelines dictate that LEO satellites may remain in orbit for 25 years following end-of-mission (EOM), while the United States Federal Communications Commission (FCC) and ESA specify a five-year deorbit period following EOM for qualifying satellites. There has been some previous work in this space, including recent papers exploring the efficacy of a variety of LEO PMD rules [3] or papers analyzing the maneuvering and technology [4] to do so, that all coalesce around the idea that disposal as quickly as possible following end of life is the preferred action for satellite operators. Currently, we do not have a clear understanding of historical PMD practices to contextualize how past missions fit into this paradigm, nor is there a thorough analysis of the multitude of factors that affect PMD timelines and viability as a means of informing what that may look like even in the current context (as 25 years have not yet elapsed since the IADC guidelines were issued) let alone in a future context, making this a critical research question at this juncture. This paper primarily uses historical RSO data from Space-Track.org, a web platform provided by the 18th Space Defense Squadron of the United States Space Force, to examine the deorbit timelines of decayed satellites across the LEO population and further interrogate the behavior of active satellites at EOM to compare PMD behavior with international and state-level guidelines and regulations. Assessments are then made regarding operators, RSO type and orbital parameters at EOM, and trends amongst constellation satellites are viewed on aggregate rather than analyzing individual objects. This paper concludes with commentary on how these trends may influence a future LEO space environment.

References: [1] European Space Agency. (2024, July 19). ESA Space Environment Report 2024. https://www.esa.int/Space_Safety/Space_Debris/ESA_Space_Environment_Report_2024 [2] Witman, D., Olson, T., Williams, B., Kesler, D., Marchand, B., “Action-Free Inverse Reinforcement Learning for Evaluating Satellite Similarity and Anomaly Detection,” Advanced Maui Optical and Space Surveillance Technologies (AMOS) Conference, Maui, HI, Sept. 17-20, 2024. https://amostech.com/TechnicalPapers/2024/Machine-Learning-for-SDA/Witman.pdf [3] Lewis, H.G., Yazadzhiyan, V., “Evaluation of low earth orbit post-mission disposal measures,” Journal of Space Safety Engineering, Volume 11, Issue 3, 2024, Pages 526-531, ISSN 2468-8967. https://doi.org/10.1016/j.jsse.2024.03.008. [4] Rhatigan, J.L., Wenschel, L., “Drag-enhancing deorbit devices for spacecraft self-disposal: A review of progress and opportunities,” Journal of Space Safety Engineering, Volume 7, Issue 3, 2020, Pages 340-344, ISSN 2468-8967. https://doi.org/10.1016/j.jsse.2020.07.026.

ABSTRACT. Knowledge sharing and integration are central to addressing 21st-century grand challenges. These function effectively through a strong global institutional network of diverse policies and frameworks. Among the global institutions that integrate the global innovation system, bilateral agreements between countries are among the most important mechanisms. In this background, the current paper explores how bilateral agreements shape the global innovation system, its actors and their networks. Any global innovation system comprises R&D by multinational firms, international technical alliances, international flows of science and technology personnel, joint international publications, international technology transfer and international trade of capital and high-technology goods. This paper specifically addresses bilateral technological collaborations, which have strengthened and amplified in the wake of increasing openness and integration of national innovation systems. An attempt is made to explore the significance and role of bilateral agreements in directing and shaping international technological collaborations and innovation. This question is answered using the case study of bilateral cooperation between the United States and India as well as the United States and Brazil. The US is one of the most influential players in international technology collaborations, with the ability to steer the direction of global cooperation. This study considers the specific technology domain of green technology and analyses the bilateral agreements signed by India-US and Brazil-US in green technology cooperation. India and Brazil are the two most important players from the developing region in the green energy sector with the capability to determine global transition to sustainability. The method used in this study is process tracing. In this method, a causal mechanism is established, a process is traced, and diagnostic evidence is gathered to prove the mechanism. This is undertaken explicitly by examining the trajectory of change and causation in bilateral agreements against the joint patent data, technology trade and FDI and international trade in goods and services complemented by empirical data collected from firms and governments in India and Brazil. The causal factors such as geopolitics, and political and economic relations are also analysed to explain the changing trends in bilateral technological collaborations between these countries. The major output of this research is the trends in bilateral technology cooperation in green energy and green technology, drivers of collaboration, and their significance in developing nations' innovative capabilities.

ABSTRACT. In the context of the platform economy, algorithmic management has greatly changed the form of traditional jobs. Work Design Theory reveals that different characteristics of a job can significantly influence worker-related outcomes. This study investigates how algorithmic management reshapes work design on ride-hailing platforms, specifically focusing on job characteristics and their influence on drivers' outcomes. Based on existing results, we find that different algorithmic management functions directly impact key job characteristics, including autonomy, skill variety, and feedback. These changes have significant effects on drivers' work outcomes at three levels. From performance level, we find a positive association between algorithmic management and efficiency, as drivers follow platform-optimized routes and schedules. From well-being level, due to algorithmic monitoring, job satisfaction may vary, while emotional labor is expected to increase as drivers navigate customer expectations and platform demands. In terms of career level, we anticipate job security to be perceived as volatile, as algorithmic performance rating impact drivers’ access to future work opportunities. The study has practical implications for a better design of platform policies to improve gig workers' well-being and societal inclusivity.

ABSTRACT. The Republic of Korea's National Science and Technology Information Service (NTIS) is a pioneering digital platform that integrates and manages national research and development (R&D) information, aimed at improving the efficiency and effectiveness of R&D investments. Launched in 2008, NTIS serves as a comprehensive resource for managing R&D data, performance evaluations, and policy coordination. By centralizing information on R&D projects, researchers, equipment, and results, it enables policymakers, researchers, and the general public to make informed decisions while reducing redundant investments across the national R&D landscape.

NTIS was developed in response to the challenges faced by the Korean government in managing rapidly growing national R&D investments. The absence of an integrated information system hindered the efficiency of government efforts to monitor, evaluate, and coordinate R&D initiatives. Prior to NTIS, various ministries operated fragmented R&D management systems, leading to inefficiencies such as data duplication and difficulty in tracking the outcomes of R&D investments. Recognizing the need for a unified approach, the Korean government, under the Ministry of Science and ICT (MSIT), initiated the NTIS project. The platform was built to provide not only a database of national R&D projects but also analytical tools that allow policymakers to assess trends, monitor project outcomes, and ensure the effective allocation of resources.

The evolution of NTIS from its pilot stage in 2006 to its current version, NTIS 5.0, illustrates its growing capacity to support Korea's R&D policy needs. NTIS 1.0 focused on the establishment of a pan-governmental cooperation system, while subsequent versions expanded its capabilities to include real-time project tracking, enhanced data analysis features, and services tailored to researchers and businesses. By 2015, NTIS had accumulated over 4.7 million data entries and attracted more than 240,000 users, with notable improvements in data accessibility, including a real-time information collection rate of 92%.

Key to NTIS’s success has been its ability to integrate information from over 15 ministries and government agencies, standardize data formats, and provide publicly accessible tools for analyzing R&D performance. This robust system has not only contributed to Korea's innovation policy but has also attracted international attention. Countries like Vietnam and Kazakhstan have sought to replicate the NTIS model, leveraging Korea's expertise in digital R&D information management.

NTIS’s expansion has been driven by increasing demands for more sophisticated analysis, including predictive analytics, trend analysis, and the visualization of research ecosystems. Recent advancements in NTIS 4.0 and 5.0 focus on making data more user-friendly, providing APIs for developers, and introducing intelligent services such as chatbots and advanced data analysis platforms. These innovations aim to make NTIS a more powerful tool for users across the entire R&D policy life cycle, from planning and budgeting to evaluation and commercialization.

The Korean experience with NTIS offers valuable lessons for emerging economies looking to enhance the effectiveness of their R&D systems. Through the establishment of a centralized platform that promotes collaboration, transparency, and data-driven decision-making, NTIS demonstrates how digital knowledge management can significantly improve national R&D policy coordination, increase research productivity, and foster innovation.

ABSTRACT. To engage SMEs in innovation could be important to economic growth in Bolivia, it is challenging to develop and achieve business model innovation due to several factor like inability to establish clear demands for technology, limited access to technological resources, insufficiently educated and trained human resources, and a lack of scientific and technological support. Also, the country's technological capabilities, market access, management assistance and other related factors play significant roles in hindering its innovation progress. This research seeks to explore and advance understanding of how SMEs innovate their business models in a lower middle-income country like Bolivia through participation in university-led cluster initiatives. To achieve the research objectives, a literature review and explorative qualitative research methodology were employed, through case studies and interviews for data collection. Empirical data was collected from the perspective of SMEs’ owners/managers, who receive external support, to examine how tailored support mechanisms can enhance SMEs business models. Based on the literature review, it was found that universities can influence insights from a diverse range of research studies to support BMI in SMEs, by structuring and designing different support activities. This can be achieved through: Facilitating knowledge or technology transfer from universities, exploring networks of relationships between universities and SMEs, SMEs seeking support to effectively manage and address problems, and the government or other institution incentivize the relationship. Each contributes to BMI by enhancing value creation, value delivery, and/or value capture. Furthermore, when examining how cluster initiatives currently impact the business model innovation of SMEs in Bolivia, the analysis revealed that university support predominantly concentrated on the development of value creation, such as technical solutions and laboratory resources, while less emphasis was placed on the dimensions of value delivery and capture. In this context the focus is on technology transfer and capacity-building; however, there is a need for more comprehensive support in value delivery and capture dimensions. Additionally, successful cases of SME business model innovation in this context, have innovated their business models in two principal ways: following a technology-driven BMI pattern, with a circular approach and technology and product development and/or innovated with a market-driven pattern, with market focusing and customer understanding and expanding customer access. Macroeconomic factors supports good access to natural resources and reliance on the informal part of the economy. In the successful cases, we found adherence to regulations and use of higher education resources as important factors for the SMEs’ enhancement of their value creation and capture processes through continuous business model innovation. In conclusion, this research elucidates the complex interplay between various stakeholders and factors influencing BMI in Bolivian SMEs. By leveraging support activities in collaborative university and other partnerships, SMEs can navigate challenges and capitalize on opportunities to drive business model innovation.

ABSTRACT. Latin America’s development continues to be a major challenge. By 2023, the decade’s average annual growth has been only 0,8%, far below the 2% of the so-called “lost decade” of the 1980s (CEPAL 2023). Empirical evidence indicates that the modest achievements are associated with low investment in Science, technology (STI) and innovation activities (Daude and Fernández-Arias 2010, Crespi and Zuniga 2012). In LAC, the average investment in R&D&I in relation to Gross Domestic Product (GDP) was 0.75% (RICYT, 2021), while the average for OECD countries is 2.4%. This historical gap in the public-private investment in innovation between LAC and OECD countries was one of the aspects that legitimized the recommendations that insisted on focusing levels of investment in R&D and innovation by means of different strategies (Arocena, 2004). One of the recommendations was the creation of Innovation agencies, as relevant institutional phenomena in countries with highly dynamic innovation systems such as Finland and Sweden. At different times in the 20th and 21st centuries, they were also present in the systems of developing countries such as Brazil, India, and Thailand. Despite this, there is still no agreed definition of what an innovation agency is, as each country has proposed an ad hoc definition based on the specific characteristics of its innovation system and the functions it performs (Foro Consultivo Científico y Tecnológico, 2010).

This work focses on this specific area of research, and aims at problematizing the science, technology and innovation (STI) governance and government in Uruguay. More specifically, it studies the evolution of the STI institutional environment, and the specific role of the executive agency of research and innovation (ANII). The paper seeks to contribute to the theoretical and empirical debate on this type of executive agencies, which are a fundamental pillar of STI policy in the Latin American region, the associated principal-agent relationships, and their consequences on the orientation and type of STI policy instruments designed and implemented in Uruguay. Questions like how were the innovation policy instruments adapted to the Uruguayan context? Have the design and implementation of these instruments undergone modifications under governments of different ideological signatures? What specific characteristics of these instruments have been shaped by the country's political regime and culture? In this sense, this proposal seeks to contribute to the discussion on how Uruguay adapted international policies and discussions on innovation to its local context.

Until recently, since the beginning of the current century, Uruguay’s science and technology environment was heavily concentrated on the single public university, Universidad de la República, a very large institution with 175 years of history and more than 150,000 students, which was complemented by the Clemente Estable Biological Research Institute (IIBCE, 1927), and reinforced by the Program for the Development of Basic Sciences (PEDECIBA, 1986) (Davyt 2012, Baptista 2016, Bortagaray 2017, Zeballos Lereté and Bianco 2021). The early ages of the STI in the 1960s, were defined by the the creation of the National Council for Scientific and Technical Research (CONICYT), as in most Latin American countries (Sagasti 2011, Feld 2020). The new century brought a new governance system, with a more complex institutional environment related to STI policy, with a new Agency of Research and Innovation (ANII), an inter-ministerial cabinet on Innovation, a new public technological university, (UTEC), the Pasteur Institute of Montevideo, and the strengthening and increase of STI policy instruments (Bianchi, Bortagaray et al. 2021). The novel institutional design kept some fragments of the previous landscape, while introducing novelties aimed at the transversality and hierarchical articulation of the policy setting (Ardanche 2012, Bianchi, Pittaluga et al. 2016, Zeballos Lereté and Bianco 2021). The new design pursued the specialization of the actions of design and regulation at the Inter-ministerial Cabinet, of deliberation at the CONICYT and of execution at the ANII (Ardanche 2012). For the Inter-ministerial cabinet to operate, it required specific political powers, to act as the principal, and the capacity to control ANII’s actions. This institutional architecture worked between 2005 and 2016. The lack of that ruling actor, in charge of defining the normative orientation of STI policy had severe effects on the STI institutional environment. To explore these implications, the paper concetrantes on a novel program, the Uruguay Innovation Hub (UIH), designed in 2022 and instrumented only this past year, for the promotion of innovation, to “propel Uruguay to the forefront of the knowledge economy” by “consolidating the local innovation ecosystem through the implementation of new instruments and the development of initiatives that foster collaboration and synergies among the various actors in the ecosystem”. Its implementation has been accompanied by a substantive change in the governance of the Agency, which has not been openly discussed or explained, but which seems to have major implications on the Agency’s autonomization from its principal and the controlling mechanisms, which in turn affects the innovation policy instrument promoted. The absence of control has been reinforced by a low political prioritization, the absence of a problem-oriented framework, the lack of a holistic innovation policy (Borrás and Edquist 2019), and the incipient emergence of actors that captured the design of innovation policy instruments in their own interest. The analysis is based on official documents and interviews with qualified informants, and a historical approach through which the different events and their implications on the governance of STI policy in the country are traced. The work concludes with some general reflections, and working hypotheses for future research agendas in the field of STI policy studies.

ABSTRACT. Background and Rationale Artificial intelligence is a field that has made another revolution in the Fourth Industrial Revolution and has become one of the most widely used emerging technologies in the present era. Further investigation of the trends in this field shows that it has shown increasing growth in recent years. Also, in the future, this trend will continue and fundamental changes will occur in societies. For these reasons, countries must definitely pay the necessary and sufficient attention to this field in their growth and development path. To achieve this growth path, it is intended to evaluate the status of leading countries in the field of artificial intelligence using two key variables in technological fields as a model for other countries. Patent is an indicator that can measure the level of innovation within a company or country. On the other hand, patents in their commercialization process require investment. In reality, there is a significant gap between capital markets and the financing of research and development (R&D) sectors. The primary reason for this, lies in the imperfect flow of information about the potential success of a technology between R&D of the firms and investors. The relationship between firms and venture capital investors is hindered due to the absence of reliable information flow. The major challenge for venture capital investors in entering a high-tech market is the information asymmetry and the high level of uncertainty stemming from the lack of familiarity with firms. Investors are unable to accurately assess the true potential of firms for innovation and their performance. Our main theory is that patents signal investors and demonstrate existing innovation. Patents have the message of novelty (inventive step) and utility of the innovation. Patents, once registered, acquire a new entity and are regarded as a category of assets belonging to firms. Investors also choose among various options based on the quantitative and qualitative status of patents. In the present study, the quantitative characteristics of patents are considered and the relationship between the number of artificial intelligence patents and attracting venture capital in this field will evaluate. Another quantitative variable that can attract investors is the number of patent applicants. A higher number of individuals or firms involved in a patent reduces the perceived risk of low quality in the eyes of investors. This indicates that more entities are invested in its success. The more connections individuals and firms have with a patent, the greater its potential customers. We also will investigate the effect of collaborative patents as a moderator variable; supposing that the more collaborative patents in the countries, the more relationship between major variables (here as: patent and VC). Collaborative patents send stronger signal to VCs, in aspect of the idea of the patent is more valuable just because of it has more than one applicant. So, our research question is: What is the impact of patent filing on the amount of venture capital acquisition? Additionally, how does this impact change when considering the effect of collaborative patents? Methodology For this purpose, we will integrate the data of number of patents and the amount of venture capital investment in some selected countries. We choose leading countries in the field of AI. Our time scope is between the years of 2012 and 2022; In these recent years, the growth of AI is salient and this would better guide us to make implications. We use artificial intelligence cooperative patent classification (CPC) technology classes of patents in Lens.org database to finding all AI patents. For collaborative patents, we look at the applicants of patents, if there are multiple applicants from single country this is a domestic collaborative paten; and if the applicants are from different countries, this patent is an international or foreign collaborative one. The fixed-effect regression model will use to measure the constant effects between the main variables. Also, the variable of the number of collaborative patents will consider as a moderating factor for the effect of the number of patents on attracting venture capital. In order to controlling the endogeneity in major variables, we want to examine the effects under the condition of time lag effect of independent variable. We also, use time lag effect of dependent variable in a two-step system GMM for Robustness analysis. Anticipated Results Ultimately, we suppose to have positive and significant relationship between the number of patents and the amount of venture capital attracted in different countries. We also, aim to evaluate various scenarios of the impact of collaborative patents including the presence or absence of collaborative patents, number of collaborative patents, and number of foreign collaborative patents on the relationship of the main variables. At first glance, it seems that having collaborative patents will lead to positive and significant responses. Artificial intelligence as one of the key pillars of the Fourth Industrial Revolution, will be extensively utilized in the process of digitalization in the coming years. Recent advancements in the growth of patent numbers, the widespread adoption of its innovations in various products and processes, and the increase in investments in this field all support this claim. Therefore, it’s so important to figure out the relationship between two crucial variables in innovation ecosystem in AI field. This would have great implications for policy makers to how can regulate in order to foster amount of capital invested in their countries.

ABSTRACT. Planetary protection is a set of space exploration practices that aims to prevent biological cross-contamination between different planetary bodies. Forward contamination (the subject of this study) refers to the transfer of Earth life to another planet, asteroid, or moon, while backward contamination refers to life of extraterrestrial origin being transferred to the Earth.[1] Planetary protection is supported by Articles I and IX of the United Nations Outer Space Treaty, which state that celestial bodies are the province of all humankind and should be explored in a manner that avoids harmful contamination and adverse environmental changes.[2] The Committee on Space Research (COSPAR) categorizes missions based on astrobiological potential of the destination, mission architecture, and whether samples would be returned to Earth. Individual nations are responsible for ensuring that their own space organizations follow these guidelines. As interplanetary missions increase in frequency and complexity, planetary protection would benefit from a scheme that classifies types of forward contaminants. This can be based on the contaminants’ biological, chemical, and logistical features, and can help actors across industry, academia, and government ensure compliance with international agreements.

Historical space events have sparked discussion about what constitutes problematic forward contamination. Examples include the SpaceX launch of a presumably unsterilized Tesla towards Mars, the lunar crash-landing of Israel’s Beresheet lander which contained biological samples, and even NASA’s disposal of astronaut waste on the lunar surface during the Apollo missions.[3,4,5] Biological material is constantly being transported into space, but it is not always problematic. Complete spacecraft sterilization is not feasible, and specific mission architectures may even integrate biological components. Forward contamination becomes problematic when the biological material has a strong chance of negatively affecting future exploration.

COSPAR’s categorization does not explicitly consider the different types of biological contaminants that may travel on a spacecraft. This may lead to ambiguities in planetary protection enforcement as robotic and crewed space missions introduce novel biological elements. The development of a lexicon that explores these different types of forward contamination can be a useful tool for spacefaring organizations to refer to through the stages of mission design.

To this end, a literature review was conducted to explore which biological components have been included in previous missions, assess how different contaminants may significantly alter extraterrestrial environments, and explore key case studies of historical forward contamination events. This information was organized into a scheme that conveys the different types of forward contaminants, their associated risks, and whether each type of contaminant has been relevant to previous space missions. This synthesis can provide users with a guideline as they create space missions with biological components, be it microbes that survive spacecraft cleaning, biological reagents for experiments, or even astronauts. The landscape of spacefaring organizations is becoming increasingly diverse and individual nations are making decisions about how to integrate planetary protection into regulations. This project aims to contribute to entry-level literature, clearly explaining how different biological elements may affect space environments.

1. “Planetary Protection,” NASA, July 28, 2024, https://sma.nasa.gov/sma-disciplines/planetary-protection. 2. “Treaty on Principles Governing the Activities of States in the Exploration and Use of Outer Space, including the Moon and Other Celestial Bodies,” United Nations Office for Outer Space Affairs, January 27, 1967, https://www.unoosa.org/oosa/en/ourwork/spacelaw/treaties/introouterspacetreaty.html. 3. Keith A. Spencer, “Violating international treaties, Elon Musk's space Tesla doesn't seem to have been sterilized,” Salon, February 12, 2018, https://www.salon.com/2018/02/12/why-sending-a-tesla-into-orbit-is-a-slap-in-the-face-to-science/. 4. Brian Resnick, “Apollo astronauts left their poop on the moon. We gotta go back for that shit.,” Vox, July 12, 2019, https://www.vox.com/science-and-health/2019/3/22/18236125/apollo-moon-poop-mars-science. 5. Daniel Oberhaus, “A Crashed Israeli Lunar Lander Spilled Tardigrades on the Moon,” Wired, August 5, 2019, https://www.wired.com/story/a-crashed-israeli-lunar-lander-spilled-tardigrades-on-the-moon/.

ABSTRACT. The implementation of sustainable development (SD) has been gaining traction among organizations, communities, scholars, policy makers, regions, and nations. The literature does, in fact, agree that innovation-centered techniques should be used to address sustainability. Therefore, the new Agenda 2030 for Sustainable Development is built on the seventeen Sustainable Development Goals (SDGs) and associated 169 objectives (sub-goals). They fairly consider the economic, social and environmental objectives of sustainable development (Dhahri and Omri, 2018). However, to achieve these objectives will necessitate taking steps or measures in a number of areas such as exploring and optimizing the potential of technological innovation (Cancino et al, 2018). Globalization has made easier for developed nations to transfer new technology to developing countries. Though, this spread of technology innovation has not often translated into actual real economic growth due to lack of technological capacity and ability of countries to promote innovative structures. However, numerous scholars have already shown the importance of technological innovations in attaining sustainability (Divella and Sterlacchini, 2021; Su and Moaniba, 2017; Hu, 2020) Despite the interesting role played by technological innovation in fostering a sustainable society. The existing literature on the theoretical and methods used, have some gaps. Theoretically, most scholars focused on the impact on the technological innovation on each aspect on the three pillars of the sustainable development, such economic growth, preserving the environment and social progress (Ahmad, et al. 2020; Shen. et al. 2021, Benn et al. 2021, Raihan, et al. 2022; Chen, et al. 2023; Sharif, et al. 2023). Few studies have experimentally investigated the impact of the technology innovation simultaneously affecting the three pillars of sustainable development. A study by Ormi, (2020) investigated the ability of the technological innovation of simultaneously affecting economic growth, social progress and environmental quality in 75 low, middle and high income countries by demonstrating the effect of the developmental stages of countries. The findings revealed that technological innovation contributed simultaneously to the three pillars of sustainable development in the long-run but only for developed economies and not effect is observed for middle and low income countries. However, this paper did not take into account of the heterogeneous effect of the technological innovation on the sustainable development in these countries. Methodically, the impact of technological innovation on sustainable development across climate regimes has not been investigated as well. Climate regimes, following Magazzino et al (2021) and Espoir. et al (2022), is understood as given the variety of climate and temperature in one land area (in one country), the primary factors used to classify countries into one given regime are the dominance of a particular climate in that specific land area. In recent years, innovation in climate mitigation technologies has seen significant progress. The stage of the climate change tends to push more works in order to adapt and mitigate its effect. Sharing the insights of the level the technological innovation to promote a sustainable society might be crucial for policy makers, scholars, industry. With regards to these gaps, the aim of this study entails to investigate the ability of technological innovation to simultaneously promote economic growth and human development, and reduce the effect of climate change in Africa across climate regimes. Two hypotheses are therefore derived: first hypothesis assumes that technological innovation simultaneously increases economic growth and human development, whereas the second assumes the effect differs across climate regimes In order to close this gap, we consider 6 African countries (i.e Egypt, Mauritius, Nigeria, Tunisia, Uganda, and South Africa) for the period 1990-2020 due to data availability, one country per climate regime is selected (like in Magazzino et al. 2021). This is due to the fact that most countries do not have information on the technological innovation index. Using Espoir et al (2022), the continent of Africa can be classified into six climatic regimes (subtropical moist (STM), subtropical dry (STD), warm temperature moist (WTM), tropical moist (TM), tropical dry (TDR), and tropical desert (TDS). We collected data Gross domestic product per capita, CO2 emissions, Human development index, Technological innovation and governance index. Findings revealed that technology innovation contributes simultaneously to the three pillars of sustainable development but the effect is heterogeneous in Africa in the long-run. The results reveal also that climate change increases the demand of technological innovation while the quality of governance has detrimental effect. Our extended analysis based on climate regimes, revealed that technological innovation and others covariates are affecting each other in heterogeneous fashion in each country. Climate change increases the demand for technological innovation in Tunisia in regime STM, South-Africa in regime WTM and Mauritius in regime TM. Future research directions policy and implication are also discussed.

ABSTRACT. Undergraduate researcher for the poster presentation:

Low-Earth orbit (LEO) has seen unprecedented growth in the number of satellites reaching orbit over the past decade. This increase in space activity is largely driven by the deployment of large satellite constellations, ridesharing, and space commercialization. [1] In addition to operational assets, orbital debris such as rocket bodies, metal shards, remnant pieces of spacecraft, and defunct satellites pose a significant threat to the continued operation of satellites in LEO. With the rapid growth and congestion of LEO space activity, satellite operators in this environment are experiencing increasingly closer conjunctions, meaning satellites and orbital debris are passing closer to one another with more frequency. [2] Current conjunction risk assessments are provided mainly by the Combined Joint Force Space Component Commander (CJFSCC) and the 18th Space Defense Squadron (18 SDS) through space situational awareness (SSA) services.[3] These conjunction warnings, given in the form of conjunction data messages (CDMs), contain valuable information for satellite operators including basic satellite information (name, type, etc.), time of closet approach, collision probability, among other metrics which are outlined in the Conjunction Data Message Blue Book.[4] Apart from satellite operators using individual CDMs to inform about potential high-risk conjunctions, CDMs have not been extensively analyzed in aggregate, particularly in a historical context. Some research reports including using CDMs to analyze minimum maneuverability requirements, analyzing CDM covariance consistency, and predicting future CDM metrics have been developed.[5] However, there is more to learn from CDMs such as determining which operators and orbital regimes are most affected by this growing risk which can help in prioritizing space traffic management (STM) goals.

Proposed STM policies including those developed by Aerospace Corporation Center for Space Policy and Strategy offer few recommendations for how to act during high-risk conjunction events that may require a satellite maneuver.[6] Without clear decision-making norms, operators face uncertainty when responding to conjunction warnings which risk continued satellite operations and future space sustainability in LEO. This means that there is likely an unequal burden on satellite operators utilizing orbital regimes that are more congested and therefore incur more frequent and higher-risk conjunctions.

This study will examine trends found in historical CDMs to determine whether certain orbital regimes and satellite operators are disproportionately affected. CDMs provided by the Space Force through an Orbital Data Request will be scraped, and based on the specific CDM metrics, conjunction data points will be organized into groups of operator type and orbital regime and compared based on collision probabilities and frequency of high-risk conjunctions. Based on the insights provided from CDM analysis, emerging risk patterns will be determined for satellite operators and orbital regimes. Given these risk patterns which are hypothesized to exist unequally in LEO, this study will provide insight on how collision avoidance goals can be met through policy development to enhance the safety and sustainability of satellite operations.

This research highlights the importance of understanding how different risk levels in LEO conjunction events impact space coordination, sustainability, and security. As mega satellite constellations and large numbers of orbital debris permeate LEO, it is critical to mitigate collision risk and orbital debris creation by developing norms among satellite operators when dealing with high-risk conjunction events.

[1] Gian Luigi Somma, Hugh G. Lewis, and Camilla Colombo, Space Debris: Analysis of a Large Constellation at 1200 km Altitude (paper presented at the 69th International Astronautical Congress, Bremen, Germany, October 1–5, 2018), International Astronautical Federation (IAF), November 2018.

[2] NASA, NASA Policies and Guidance for Conjunction Assessment and Mitigation (October 2023), https://www.nasa.gov/wp-content/uploads/2023/10/conjunction-assessment-and-npr-8079.1.pdf.

[3] https://www.space-track.org/documentation#/odr

[4] The Consultative Committee for Space Data Systems, Conjunction Data Message Recommended Standard, CCSDS 508.0-B-1, Blue Book (June 2013), https://public.ccsds.org/Pubs/508x0b1e2c2.pdf.

[5] Daniel Moomey, “A Framework for Minimum Maneuverability Requirements for Low Earth Orbit Conjunction Assessment, Using Historical Conjunction Data Messages,” Journal of Space Safety Engineering 9, no. 3 (September 2022): 375–384, https://www.sciencedirect.com/science/article/pii/S246889672200043X#sec0001; B. Reihs et al., “Analysis of CDM Covariance Consistency in Operational Collision Avoidance,” in Proceedings of the 7th European Conference on Space Debris, ed. T. Flohrer and F. Schmitz (Darmstadt, Germany: ESA Space Debris Office, June 2017), http://spacedebris2017.sdo.esoc.esa.int/proceedings/sdc7/paper/449/SDC7-paper449.pdf; Francisco M. Caldas, Cláudia Soares, Cláudia Nunes, and Marta Guimarães, “Conjunction Data Messages for Space Collision Avoidance Behave as a Poisson Process,” March 27, 2023, https://arxiv.org/pdf/2303.15074.

[6] Michael P. Gleason and Travis Cottom, U.S. Space Traffic Management: Best Practices, Guidelines, Standards, and International Considerations (El Segundo, CA: The Aerospace Corporation, Center for Space Policy and Strategy, August 2018).

ABSTRACT. INTRODUCTION Society's concern over climate change has intensified, particularly after IPCC reports urged action to mitigate human-induced impacts. While governments and IGOs have traditionally been held accountable, multinational corporations (MNCs) also bear significant responsibility due to their polluting activities and broader environmental impacts. This research explores how MNCs are addressing climate change, focusing on SDG 13 (Climate Action).

THEORETICAL BACKGROUND Studies examine the relationship between MNCs and climate change in various contexts. Kolk and Levy (2003) found that U.S. and European MNCs, particularly in oil and automotive industries, initially adopted distinct strategies influenced by institutional environments and business-government norms. Over time, regulatory frameworks drove strategic alignment. However, MNCs previously investing in renewables remain hesitant to reinvest due to past losses.

Recent research highlights the need for MNCs to balance global standardization with local adaptation and better integrate sustainability into corporate strategies (Kolk, Kourula & Pisani, 2017). Despite progress, gaps persist in understanding corporate adaptation to climate risks, necessitating clearer evaluation frameworks and systematic data collection (Averchenkova et al., 2016). As geopolitical tensions rise, MNCs face increased pressure to adopt regenerative sustainability practices (George & Schillebeeckx, 2022).

METHOD AND ANALYSIS Using bibliometric analysis, this study examines patterns in research addressing MNCs and climate change. A search on Dimensions yielded 73,488 documents, narrowed to 543 relevant to SDG 13. These include articles, chapters, preprints, and more. VOSviewer software visualized bibliometric networks.

Most documents fall under "Commerce, Management, Tourism and Services" and "Strategy, Management, and Organizational Behavior," with a noticeable increase in publications following key IPCC reports. Authors from developed countries dominate the sample; of the 11 most prolific researchers, only three are from emerging economies (China and Brazil). Similarly, most institutions contributing to this research are based in developed countries, with Brazil’s University of São Paulo as the sole exception among emerging economies. Co-authorship links reveal stronger collaborations between institutions in developed countries or between developed and emerging economies, with limited connections solely between emerging nations.

IMPLICATIONS AND PRELIMINARY RESULTS This research underscores the growing role of MNCs in addressing climate change. The analysis shows that most involved researchers, institutions, and countries are from the developed world, reflecting the concentration of MNCs there. China stands out for its efforts to address pollution and climate issues, while Brazil’s prominence stems from its critical role in hosting the Amazon rainforest, a key player in global ecological balance.

Despite increased collaboration, partnerships tend to align with socioeconomic profiles, leaving room for greater cooperation among emerging economies. Moving forward, this study will analyze specific MNC activities to assess their impact on climate change mitigation.

ABSTRACT. Advancements in generative artificial intelligence (GenAI) have given rise to widespread enthusiasm towards the application of GenAI tools in academia settings among educators and students alike. While GenAI offers promising opportunities to transform the future of higher education and the scientific enterprise (Rawas, 2023; Rudolph et al., 2023; von Krogh et al., 2023), the incorporation of GenAI systems also raises potential risks to academic integrity, privacy, and equity (Eke, 2023; Lund et al., 2023). In response to concerns over these potential risks, governments and higher education institutions worldwide are developing regulatory frameworks to stay abreast of technological advancement in GenAI and ensure its responsible use (McDonald et al., 2024; Smuha, 2021).

Academic scientists serve multiple roles in the development, application and governance of GenAI. However, their uptake of and attitudes towards the regulation of GenAI are less documented. In this paper, we draw on data from a 2023 survey of university-based scientists in six disciplines—biology, chemistry, computer and information science engineering, civil and environmental engineering, geography and public health—from a nationally representative sample to examine the underlying factors driving their use of and attitudes towards the regulation of GenAI. The extensive literature on technology adoption has proposed a broad range of theories, including the Theory of Planned Behavior (TPB), the Technology Acceptance Model (TAM), the Innovation Diffusion Theory (IDT) and the Unified Theory of Acceptance and Use of Technology (UTAUT) (Ajzen, 2011; Davis et al., 1989; Fishbein & Ajzen, 1977; Rogers, 1995; Venkatesh et al., 2003). These theories collectively identified several clusters of determinants to technology acceptance, including perceived benefits, perceived costs, perceived ease of use and facilitating conditions. Recent research also shows that risk perceptions over technologies are important factors driving attitudes toward AI use and governance (O’Shaughnessy et al., 2022).

In this paper, we focus on two sets of factors, perceived cost over GenAI and perceived self-efficacy to make hypotheses on their relationship with GenAI adoption and attitudes toward its regulation. We conceptualize concerns over GenAI as a form of perceived cost and measure it with concerns over privacy and concerns over agency loss; and measure perceived ease of use with confidence in teaching and research. Using structural equation modelling, we find that 1) confidence in using GenAI in teaching and research is positively related to the use of GenAI in teaching and research, 2) risk perception over privacy and security is negatively related to GenAI use, and 3) risk perception over loss of agency is negatively related to favor of federal regulation of GenAI. There is no evidence that GenAI use is predictive of attitudes toward federal regulation of GenAI. Our findings offer empirical evidence highlighting the negative impact of concerns on GenAI adoption and provide insights for GenAI developers, users and regulators to promote its responsible use and governance.

ABSTRACT. This research examines how community-led dispute resolution transforms governance in digital platforms, demonstrating new forms of civil society organization in the digital age. Platforms like Xiaofating on Xianyu enable community members to act as volunteer judges in resolving buyer-seller disputes, creating a self-organized decision-making model. Similar community-driven systems exist globally, such as eBay's Resolution Center and Airbnb's Resolution Center, where users resolve conflicts through peer feedback and dialogue before involving platform administrators (Hirsch & Leclerc, 2021). The research addresses three key questions: (1) What factors motivate community members to voluntarily act as judges? (2) How do theories of social capital, collective intelligence, and altruism explain community participation? (3) What are the implications for sustainable community self-organization in digital platforms? The theoretical framework integrates several organizational behavior theories. Social exchange theory (Blau, 1964) explains how community members engage based on reciprocity and social capital building. Social identity theory (Tajfel & Turner, 1979) suggests that participants are driven by community belonging and collective responsibility. Collective intelligence theory (Malone, Laubacher, & Dellarocas, 2010) demonstrates how community-based decision-making leverages crowd wisdom. Additionally, prosocial behavior theory (Batson, 1991) explains altruistic motivations driving civic participation. Using survey-based qualitative analysis with thematic analysis (Braun & Clarke, 2006), preliminary findings indicate that social capital development and civic altruism significantly drive participation. Community members feel personal responsibility for maintaining fairness, supporting social exchange theory's prediction of non-monetary rewards through enhanced community trust (Blau, 1964). The study reveals challenges in sustaining community engagement, including decision fatigue and unclear guidelines. These findings have significant implications for supporting sustainable community self-organization in digital spaces, suggesting that fostering social capital and community belonging drives continued civic participation. The results underscore how digital platforms can enable effective community self-organization through informal governance structures where decentralized decision-making relies on voluntary civic contributions (Benkler, 2006). References: Batson, C. D. (1991). The altruism question: Toward a social-psychological answer. Erlbaum. Benkler, Y. (2006). The wealth of networks: How social production transforms markets and freedom. Yale University Press. Blau, P. M. (1964). Exchange and power in social life. Wiley. Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77-101. Hirsch, P. M., & Leclerc, R. (2021). The rise of peer-mediated systems in the digital economy: A review of Airbnb and eBay. Journal of Digital Platforms, 14(3), 45-60. Malone, T. W., Laubacher, R., & Dellarocas, C. (2010). The collective intelligence genome. MIT Sloan Management Review, 51(3), 21-31. Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33-47). Brooks/Cole.

ABSTRACT. This paper analyses the global policy shifts in human gene editing precipitated by the ethical, legal, and societal uproar following He Jiankui's controversial experiments in 2018. These experiments, involving the genetic modification of human embryos, not only revealed deep flaws in the international regulatory framework but also ignited a global debate over the implications of human germline editing. Employing a comparative approach, this study examines the responses from key jurisdictions—China, the United States, the European Union, the United Kingdom, Canada, and India—highlighting the diversity in regulatory adaptations ranging from strict enforcement to more supportive stances towards scientific advancements. Through a comparative analysis of the regulatory histories of cloning and in vitro fertilization (IVF), this paper articulates lessons and strategies for the responsible oversight of gene editing technologies. Emphasizing a stakeholder-centric approach, it advocates for a policy environment informed by empirical evidence and responsive to public concerns. The paper underscores the need for an anticipatory and inclusive regulatory framework that integrates robust ethical considerations and actively engages stakeholders, thereby aligning technological advances with societal values and ensuring ethical integrity and public trust in gene editing advancements. This approach advocates for dynamic governance mechanisms that are responsive to evolving scientific landscapes and public sensibilities, striving to balance innovation with ethical responsibility.

ABSTRACT. The platform economy has fundamentally transformed labor markets worldwide, challenging traditional regulatory frameworks and employment relationships. This study examines how worker heterogeneity in China's ride-hailing sector—the world's largest with 5.406 million drivers—influences regulatory preferences and platform efficiency. While existing policies since 2016 have sought to protect platform workers, the prevailing universal protection approach overlooks critical heterogeneity in how workers engage with platforms, particularly the distinct needs of full-time versus part-time drivers. Through a methodologically rigorous approach combining institutional ethnography and discrete choice experiments (DCE), this research investigates how employment patterns influence drivers' policy preferences across five key institutional dimensions: social insurance coverage, working hour limits, collective bargaining rights, vehicle licensing requirements, and safety training certification. The study employs a three-stage process: preliminary qualitative interviews with drivers, a pilot survey of 46 drivers stratified by employment status (43.38% full-time, 56.62% part-time), and a main DCE study with a D-optimal design comprising 24 choice sets. Preliminary findings reveal several counterintuitive patterns. Enhanced labor protections may paradoxically disadvantage part-time drivers by imposing inflexible requirements that conflict with their supplementary income-earning patterns. Full-time drivers demonstrate unexpected preferences for operational autonomy over comprehensive social protection, challenging fundamental assumptions about the relationship between employment intensity and desired regulatory protections. These insights suggest that current one-size-fits-all approaches may reduce rather than enhance platform efficiency. This research advances both theoretical understanding and practical approaches to platform regulation. It extends institutional theory by demonstrating how worker heterogeneity fundamentally challenges assumptions about regulatory legitimacy in platform markets, while contributing to labor market segmentation theory by revealing how different forms of platform engagement create divergent regulatory needs. The findings provide empirically-grounded insights for developing more sophisticated regulatory frameworks that can accommodate worker heterogeneity while maintaining market efficiency, with implications for platform economy regulation globally.

ABSTRACT. Since the advent of ChatGPT in late 2022, artificial intelligence (AI) systems have seemingly taken the general public by storm, and policymakers are no exception. Recent data from the nonprofit lobbying research group, OpenSecrets, uncovered that the total number of organizations lobbying on AI issues has gone from 158 organizations in 2022 to over 460 organizations in 2024, nearly tripling in the span of two years. This is a clear indication that interest groups are vying to be heard by politicians and policymakers. Understanding the methods used by these groups will aid in better understanding the forthcoming policy decisions taken by the federal government. In other words, in line with previous literature on lobbying, the ways in which lobbying actors present the impacts of an issue can have a leading effect of changing policy along those lines. In our case, this is the nascent policy arena of AI. Thus, this research aims to improve our understanding of the role interest groups play in shaping issues related to AI. By asking the question, how are issues framed by different lobbyist groups in Canada, this research uncovers the complex power relationships that exist between interest groups and AI legislators. This research conducts a qualitative critical discourse analysis (CDA) of 18 witness hearings held during Canadian federal parliamentary committee meetings related to AI. It is further informed through a theoretical framework that combines both CDA with Agenda-Setting Theory, allowing my observations to be grounded within theories of public administration. By taking a CDA approach to the dataset, this allows us to critically engage with the discourse and uncover the different power dynamics, discursive methods, and issue frames within different AI interest groups. In an emerging policy field such as AI, where policymakers have limited knowledge and industry holds exclusive knowledge, we hypothesize that interest groups act as “gatekeepers of information”, disseminating information according to their respective goals. This reality is especially apparent within sectional groups, where lobbyists tend to dismiss facts that would otherwise hurt their own interests. For example, the tech lobbyists testifying overwhelmingly spoke to the various economic and societal benefits brought about by AI, whilst discounting the technology’s ethical and displacement concerns. We also observe that witnesses from academia take a more impartial stance, acting as a natural “control group” in this setting. Related to lack of representation in public forums, our observations fall in line with previous literature on Canadian tech lobbying, noting that certain groups are often left to the margins of decision-making consultations. These groups include digital rights organizations, public interest advocacy groups, and Indigenous voices. Our findings hope to serve as a tool for policymakers to recognize the dynamics that lobbying has on AI policy. In recognizing that policymakers must often engage with AI lobbyists for expertise – leading to these interest groups acting as gatekeepers of information – legislators should be aware of the power imbalance that exists in the industry.

ABSTRACT. Since the early 2000s, Sub-Saharan African countries, including Zimbabwe, have experienced significant growth in internet and mobile phone usage, spurring the development of mobile applications aimed at smallholder farmers. These applications are widely recognized for their potential to transform smallholder farming into a profitable enterprise, boost food production, and promote sustainable agricultural practices. Despite this promise, sustained adoption remains a critical challenge, particularly among women farmers, who form the backbone of smallholder agriculture in the region. This study investigates the user-specific factors influencing the adoption of mobile agricultural applications, with a focus on women farmers in Zimbabwe. Guided by the Technology Acceptance Model (TAM), this research addresses three key questions: How does perceived usefulness impact the adoption and sustained use of mobile agricultural applications among women smallholder farmers? To what extent does perceived ease of use drive initial adoption and long-term utilisation of these applications? Which TAM construct-perceived usefulness or perceived ease of use plays a more significant role in shaping adoption behaviour? A cross-sectional design will be employed, targeting 200 women farmers in the Mushandike Irrigation Area, Masvingo Province. Participants will be selected through stratified random sampling to capture diverse adoption experiences. Data will be gathered via structured interviews and open-ended questionnaires and analysed using logistic regression in Stata. The findings will provide a nuanced understanding of the factors driving or hindering the adoption of mobile agricultural applications in Zimbabwe. By identifying key user-specific challenges and successes, this study offers actionable insights for developers and policymakers to enhance the design and implementation of these applications. Ultimately, the research aims to support increased adoption and sustained usage of mobile agricultural technologies, contributing to more sustainable farming practices and improved rural livelihoods across Sub-Saharan Africa.