ABSTRACT. This research explores how the saliency of matchmaking algorithms—the extent to which users think about them—affects user experiences in competitive online gaming. Employing an exploratory sequential mixed-methods design, the study begins with semi-structured interviews to examine users’ perceptions, resulting in a process model that is tested in a quasi-field survey utilizing a retrospective recall design. The findings reveal that the saliency of matchmaking algorithms, regardless of their actual interference, influences users’ gameplay experiences. A negative relationship is observed between algorithm saliency and gameplay enjoyment, mediated by internal attribution of outcomes when users lose, but not when they win. As the first study to investigate matchmaking algorithm saliency, this research identifies it as a critical factor that can diminish user enjoyment and extends understanding of algorithmic perceptions by demonstrating that mere awareness of the algorithm affects user experience. Practical implications for game developers are provided.

ABSTRACT. The research investigates stakeholder identification, selection, and involvement in digital economy policy development and implementation, focusing on Nigeria's National Digital Economy Policy and Strategy (2020- 2030). It addresses the gap in effective stakeholder engagement in developing digital policies in sovereign developing countries. The study employs the Critical Systems Heuristics (CSH) framework to analyse stakeholder roles and involvement. The overarching aim of the PhD thesis is to develop a framework for beneficiary-focused policies applicable in developing nations – given their distinct power dynamics characteristics that often question the motivation for acting on behalf of the public. Through interviews with policymakers, implementers, and beneficiaries, the research explores the rationale behind stakeholder selection, power dynamics, expertise utilisation, and perceptions of legitimacy. The findings will contribute to a new framework incorporating evidence-backed criteria for identification, selection, and involvement, thereby offering guidelines for effective stakeholder engagement in developing national digital economy policies.

ABSTRACT. Introduction Climate change poses significant societal challenges today, particularly in developing countries. It threatens progress made on Sustainable Development Goals (SDGs), including goals on health (SDG3) and food security (SDG2), affecting health outcomes of individuals and communities (World Meteorological Organization (WMO) et al., 2023). At the centre of this is climate health data, a research object in information systems (Aaltonen & Stelmaszak, 2023, 2024) and the relationship between climate data and health outcomes. Climate health data contextualise climate data in the health domain to benefit human health outcomes, and they are essential for measuring risk, adaptation, and mitigation of climate change (Findlater et al., 2021; UNCTAD, 2024; WHO/WMO, 2023). Climate health data is critical to understanding how the relationship between climate and health has behaved in the past and how it will behave in the future to support societal decision-making from global to local scales (Griggs et al., 2020; Overpeck et al., 2011). In this PhD project, I focus on climate data on food security in Malawi, which has a bearing on malnutrition in developing countries, accounting for 38 per cent of the infant mortality rate (UNICEF, 2018). Climate change affects not just food production but also food's nutritional quality (Lauro & Visioli, 2021; Owino et al., 2022). Despite the availability of climate and weather data and the importance of this in decision-making, the actionability of climate data in climate-sensitive environments remains a challenge (WHO/WMO, 2023). In addition to this, despite a mutual sense of urgency on the implications of climate change, countries differ in the breadth and organisation of the value chain, moving from raw climate data processing to the provision of actionable data (Fischer et al., 2024). This has a profound effect on how climate health data influence health outcomes (Philip Osei et al., 2022).

Context The study will be conducted in Malawi, a country in southern Africa where agriculture is the cornerstone for food security, employment, and the country’s gross domestic product (Ministry of Agriculture, 2021; NSO, 2008). The country's food security system is dependent on smallholder farmers and rainfed agriculture, making the food security system susceptible to climate change (Government of Malawi. Malawi Vision 2063., 2020; Ministry of Agriculture, 2021). The Ministry of Agriculture introduced the National Agriculture Management Information System (NAMIS) to collect, manage, and analyse agriculture and other data relevant to agriculture and food security, such as climate and weather. NAMIS is a critical foundation for climate-resilient agricultural practices and food security in Malawi. The system collects agriculture data, market data and climate and weather data, among others, to inform the food situation and security in the country. One key stakeholder in the system is the agriculture extension development officer (AEDO), who is the primary advisor to the farmer and a primary data collector for the system. They form a key link between the farmers and NAMIS.

Aims and methodology This project aims to answer the question of how climate health data can be made actionable for smallholder farmers to improve food security. Actionable data in this project is conceived as data that is correct, available, timely, understandable, and contextualised to address a real-world problem (Bankowitz, 2010; Latifov & Sahay, 2012; Leonelli & Tempini, 2020; Ramsden, 2020; Sarkar, 2022). To address these questions, the study will use action design research (ADR), which will allow intervention by involving AEDO and other stakeholders to express their needs and be involved in coming up with a solution that addresses their needs while contributing to knowledge and practice (Sein et al., 2011). The concepts of data journeys (Leonelli & Tempini, 2020) and data commodities (Aaltonen et al., 2021) will be the theoretical tools for the study. ADR and the theoretical tools will help in understanding the current practises, the intervention of the data processes, and the evaluation of the interventions to share and formalise learning.

Data journeys provide a framework that evaluates how data is shaped as it travels within social, political, cultural and technological environments (Leonelli & Tempini, 2020). The nature and conditions in which data is generated and how it travels profoundly affect how it can be operationalised and legitimised as evidence (Leonelli, 2020; Priego & Wareham, 2024). Different sites in the data journey have their norms, protocols and standards for handling the data (Edwards et al., 2011). Data journeys provide a means to understand data processes, how data is legitimised to save as evidence (Pousti et al., 2023) and give a window into how data actionability is embedded or inhibited in the data journey process (Cambrosio et al., 2020; Ramsden, 2020). In addition to data journeys, the concept of data commodities will assist in understanding how data gains value during this process of generation, management and is ensembled into tradable goods that have value for the intended user (Aaltonen et al., 2021). We conceptualise climate health data as tradable commodities that have the value of monitoring and improving food security, while the data journeys view focuses on the processes of data from generation to utilisation, which will help identify pain points in data processes throughout the data journey, enabling targeted improvements to enhance the actionability of the data.

Potential contribution We aim to contribute to the literature centred on the role of technology in shaping data and meaning across contexts (Aaltonen et al., 2023). We also contribute to the understanding of actionable data based on the end users' needs, capabilities, and modalities. Lastly, it will illuminate the information systems and data studies literature with empirical and practical contributions on developing data that are co-created with multiple stakeholders and actionable at sub-district levels within a climate-sensitive, low-resource context.

ABSTRACT. This paper describes a study aimed at exploring how agency develops when humans interact with AI, focusing on autonomous driving systems (ADS). ADS is considered one of the most influential application sectors for AI research primarily due to its advanced capability of perceiving, learning, and acting on behalf of humans, and also because it navigates highly uncertain environments that allow for low margin for error. While scholars have offered insights into autonomous driving, little is known about how agency emerges between humans and ADS. This is a challenge that extends beyond ADS to the broader AI research. In this study, an integrative lens is constructed based on the engagement between Cultural-historical Activity Theory and Material Engagement Theory. The integrative dynamics between humans and ADS are explored by drawing upon simulation-based data and real-world driving experiences. The findings provide critical theoretical enrichment to cultural-historical and material engagement explanations of agency development and offer insights into the properties of the human-AI hybrid and the complexity of autonomous systems.

ABSTRACT. This project constitutes an exploratory study, which explores and clarifies the users’ awareness about data disclosures in Metaverse and associated risks, as well as the actions that they take in order to be protected. Analysing users’ behaviour in Metaverse regarding their data disclosures constitutes a research effort to make a broad focus narrower by pinpointing both the enforcement/improvement of data protective measures and key points about a common legal metaverse framework or more effective and specialized policies.

ABSTRACT. Introduction: My research project is situated at the intersection of informatics, surveillance studies, and critical media studies, with a focus on the deployment of digital surveillance systems within law enforcement contexts. The project examines how advanced surveillance technologies—shaped by digitization and artificial intelligence—reconfigure power, governance, and individual rights.

ABSTRACT. 1.0 Purpose This study aims to explore the relationship between algorithmic management and the socialisation behaviours of gig workers in the e-hailing industry through the lens of techno- duality and socio-materiality. It also seeks to provide appropriate references and theoretical support for improving relevant management mechanisms and promoting social accountability in platform governance. E-hailing platforms closely link E-hailing company employees and drivers through algorithms. This allows company employees to manage drivers online at scale through algorithmic functionality (Wang et al., 2018; Bi et al., 2022). At the same time, it also makes drivers continuously search for methods to cope with the ever-changing algorithmic strategies of the E-hailing platforms (Li et al., 2024). Existing literature on the algorithmic management of workers has focused primarily on the characteristics of flexibility, autonomy, and diversity that people gain from working digitally (Wood et al., 2019; D'Cruz and Noronha, 2006). Researchers have also critically discussed the power inequalities and pressures that algorithmic control and surveillance bring to workers (Just and Latzer, 2017; Shapiro, 2018). However, these studies largely overlook the potential impacts of E-hailing platform algorithms after creating inequities in the driver population and how driver and employee behaviours 2 change due to using E-hailing algorithms (Bucher et al., 2021). Therefore, this research intends to explore the use of algorithms on E-hailing platforms and explicate the “roles” that such algorithms play between related workers and drivers of E-hailing companies. Based on this, the study will discuss the socialisation and resistance of the driver community to the advantages and disadvantages of the algorithms' functionality. Finally, the research will also analyse how e-hailing algorithmic accountability drivers and workers can be used in the future. This will also unpack the importance of algorithmic transparency for stakeholders such as e- hailing regulators, firms, and drivers, as well as the extent to which transparency can be achieved in reality. Through these aims, this research question is: What is the impact of e-hailing algorithms on drivers and employees within the e-hailing industry? 3 2.0 Methodology For data collection, two versions of an interview protocol have been prepared after reading the relevant literature on the topics of E-hailing algorithms, algorithm management, and algorithm accountability. One version will be used to conduct interviews with E-hailing drivers and the other version will target workers related to the E-hailing industry, such as government officials and E-hailing company workers. Before the formal interviews will commence, scoping interviews were conducted with five participants with different identities related to the e-hailing industry to obtain preliminary data for the study and to aid the design of the formal interview protocols. Participants for the scoping interviews included: a government official, an employee of the e-hailing algorithm department, and three drivers. As a next step, these scoping interviews will be used to conduct preliminary data analysis and inform the qualitative data coding methodology for the first and second phases of the study afterwards. In addition to this primary data (semi-structured interviews with 25-35 drivers and related industry workers in the e-hailing industry), secondary data (e.g., forms, reports, and news) will also be collected. Qualitative data will be coded using the Gioia method (e.g., Gioia et al., 2013) through software such as NVIVO. 4 3.0 Relevance and Contribution Through the lenses of Techno-duality and Socio-materiality, this study expands understanding of the relationship between technology and society by exploring the impacts of E-hailing algorithms on E-hailing employees and drivers (Weimo Li et al., 2023; Gan et al., 2021). Specifically, this research presents three potential contributions. First, based on the duality of technology, this study will investigate the dynamic relationship between technology and actors in the social environment by summarising the roles and functions of E-hailing algorithms on the interactions between E-hailing employees and drivers. The results of this will focus on the expected influence and adjustment role of feedforward and feedback mechanisms' algorithms on drivers’ behaviour and explore the asymmetry and control mechanisms of algorithms' rights between employees and drivers of E-hailing companies. Secondly, the study intends to expose the socialised behaviour of gig workers in the face of algorithmic management by further discussing the sociomateriality perspective. Thus, this research aims to extend previous drivers' passive behaviours towards algorithms by uncovering the mechanisms of how E-hailing drivers discover and understand algorithms and how they adapt to and fight against algorithmic management (Zhang et al., 2023). This 5 research findings will discuss the interconnectedness of algorithms and social factors such as driver behaviour and organisational decision-making from a practical perspective, thus enriching the ‘Mutual Construction’ view of technology and society (Bavdaz, 2017). Finally, the study will critically discuss the issue of responsibility of different stakeholders, such as drivers and employees, towards algorithms by introducing the concept of algorithmic accountability. By examining features such as algorithmic transparency, the study has rich potential to extend beyond the design issues of the technology itself. Instead, it can focus more on accountability issues when the technology is applied at a societal level.

ABSTRACT. This study explores the transformative impact of artificial intelligence (AI) on data governance in intra- and inter-organisational contexts. Using a 2x2 matrix framework, it explores governance practices across collaboration, trust, security, and risk management, integrating socio-technical perspectives with established governance models, it provides scalable frameworks for responsible AI adoption in data-intensive industries.

ABSTRACT. Research Background Technological advancement and the demands for 21st century skills have brought about changes in teaching and learning, advocating for an education system that addresses the individual needs of learners. Personalised Learning (PL) has emerged as a transformative approach in education, emphasizing the individualization of learning experiences to enhance student outcomes. While PL has gained momentum in education settings, its application in the context of Open Distance Learning (ODL) institutions within the African continent remains underexplored. ODL institutions face unique challenges such as limited resources, outdated teaching methods, and diverse learners needs. The absence of personalization within the ODL environment has been linked to challenges such as high student dropout rates. Additionally, the current adopted e-learning platforms lack personalization, adaptability, and fail to address the diverse needs of distance learners. The existing systems are designed to support a one-size-fits-all approach and fail to meet the individual needs of learners which are unique. Research Aim The aim of this study is to design a prescriptive framework that will guide the design of PL in ODL Environment in the context of African ODL institutions. To achieve the objective of the study, this research aims to address the following question: “What components constitute a framework for designing PL in an ODL environment?”.

Significance of the Study This study will contribute to the development of a framework to guide the design of PL in ODL institutions within the African context. The framework will assist educators within the African ODL institutions and environments in personalizing their learning process to cater to the needs of distance learners and improving their learning experiences. In addition, the framework can serve as a blueprint for educational institutions that utilize e-learning to deliver personalized learning experiences in their teaching and learning.

Methods The research design for this study is based on the philosophical assumption of pragmatism, which emphasizes practicality in bringing change and action in real world-settings. The study aims to develop a framework for personalizing learning in ODL environments within the African context. It adopts a constructive epistemological stance, focusing on knowledge that leads to action and change. The study employs Design Science Research (DSR) as a research strategy and DSR Methodology to guide the research process consisting of six phases namely: problem identification, objectives and solution, design and development, demonstration, evaluation, and communication. The study context is Open University of Tanzania as source of requirements for the design and evaluation of the framework. Data collection will employ a mixed methods approach, gathering both quantitative and qualitative data through questionnaires and interviews. The framework evaluation will include both formative and summative evaluations during its development and refinement.

Expected Contribution to Knowledge The primary contribution of this study is the development of a framework (artifact) that will guide the design of PL within the context of African ODL institutions. The framework will serve as a valuable resource for educators, providing prescriptive knowledge on how to design personalised learning processes and courses within LMSs to meet the unique needs of individual distance learners, thereby enhancing overall online learning experiences. Furthermore, this study aligns with the concept of 'exaptation' knowledge, involving the adoption of existing solutions from various domains to address new challenges. The novelty of this study lies in the utilization of exaptation knowledge by developing an artifact that adapts solutions from other contexts. The framework's design will draw upon solutions found in the literature, theories, empirical studies, and strategies utilized or designed in other domains.

ABSTRACT. As organizations increasingly explore virtual and hybrid work models, the potential of the Metaverse as a platform for workplace training and development has come to the forefront of academic and industry discussions (Alcañiz et al., 2018, Koohang et al., 2023). The development of the Metaverse has been facilitated by the advancements in virtual reality (VR), augmented reality (AR), 3d rendering, haptic technology, 5G and edge computing. These innovations enable high degrees of interaction and immersion, allowing users to transition between real and simulated environments. This integration fosters more engaging and lifelike experiences, positioning the Metaverse as a transformative platform for various applications (Dwivedi et al., 2022; Mystakidis, 2022), creating unique opportunities for implementing experience-based learning strategies in corporate training programs (Hajjami and Park, 2024). The concept of the workplace is reconceptualized beyond its traditional confines, undergoing a transformation towards a more adaptable, agile, and collaborative workforce, with a broader range of solutions to accommodate different working styles such as Work from Home (WFH) and Work from Anywhere (WFA) models (Gibbs et al., 2022). It becomes imperative to understand the impacts of Metaverse in training programs and productivity, and the related strategic implications for organizations adopting it (Polyviou and Pappas, 2023).

ABSTRACT. My overall PhD research aims to investigate how (a) Control Theory and (b) Temporal Structuring interplays within open- source knowledge environments, how it is configured, how it evolves or emerge as time progresses, considering external influencing factors, in this case GAI. As such, my first goal (and paper) is to investigate the case of Stack Overflow, to address the disruptive nature of GAI in relation to control structures within these networks. By addressing three central research questions, the study investigates: (RQ1) the underlying control mechanisms and their configurations within these networks, (RQ2) the disruptive influence of GAI on existing control systems, and (RQ3) how the introduction of GAI triggers dynamic reconfigurations of control portfolios using process theorizing approaches. The study responds to growing concerns about the governance of digital communities, particularly those reliant on user-driven contributions and reputation-based systems. Understanding how control evolves in response to GAI offers actionable insights for platform designers, moderators, and policymakers navigating similar disruptions.

ABSTRACT. Industry 4.0 has provided a focus on the introduction of advanced digital technologies into business and industry, such as robotic process automation (RPA), internet of things (IoT), big data analytics, and artificial intelligence (AI). Hyperautomation utilises a confluence of these Industry 4.0 technologies to provide a focus on the automation of as much as practicable (Bornet et al., 2020). However, there is recognition that Industry 4.0 needs to develop; in Japan, Society 5.0 has emerged (Japan Business Federation, 2016), whilst in the European Union, Industry 5.0 has been conceptualised (Breque et al., 2021). Supply Chain 5.0 builds on Society 5.0 and Industry 5.0, to provide contextualisation of these philosophies to supply chains. This move towards Supply Chain 5.0 shifts the emphasis from mere automation and digitalisation, to a more sustainability focussed and human-centred approach, providing greater alignment with UNESCO’s sustainable development goals (SDGs). Contemporary research in this area largely takes a positivist stance, with limited qualitative studies on hyperautomation in the supply chain, often concentrating solely on manufacturing and overlooking logistics (Farzana et al., 2023; Fedosovsky et al., 2022). This research aims to extend prior work (Birkbeck & Rowe, 2024) to understand the present and future challenges posed by Supply Chain 5.0, allowing human centrality to be maintained throughout the adoption of hyperautomation.

ABSTRACT. Blockchain technology has enabled Initial Coin Offerings (ICOs), decentralised fundraising mechanisms that connect project developers and investors without intermediaries. While ICOs have witnessed widespread adoption due to their accessibility, they have also enabled fraudulent activity, such as pump-and-dump schemes and scams. This research positions the ICO network as a dynamic system where competing actor intentions and behaviours create tension. Tensions in ICO networks arise from the conflicting intentions and behaviours of actors—such as ethical participants seeking long-term growth, opportunistic scammers exploiting the system for quick profits, and the broader decentralised structure that both fosters trust and enables vulnerabilities—creating a dynamic interplay that shapes the network's evolution, stability, and susceptibility to fraud. By applying network theory, this research contributes a new perspective on how these tensions drive the evolution of the ICO landscape.

ABSTRACT. Data quality is an issue facing organisations that has not been given much attention in the literature (Xu, 2015). Although this is an issue facing organisations during software implementation of enterprise systems, not enough resources and time are given in order to address master data quality issues adequately from the start (Benfeldt et al., 2020). This results in poor master data quality affecting other phases of the software implementation without master data experts using adequate tools to address the issues (Guerra-García et al., 2023). The task of addressing master data quality issues is often avoided as the operational staff and IT staff do not hold all the knowledge on their own nor do they collaborate in order to understand these issues (Elragal & Elgendy, 2024).

A systematic literature review of 56 IS papers, using an eight-step methodology for conducting a stand-alone literature review (Okoli, 2015), over the last ten years was conducted using through the lens of the Work Systems Framework (Alter, 2013). The literature review provided a holistic organisational perspective of master data quality issues, but further research was needed to identify generative mechanisms causing issues of master data quality. The research question is: What leads to master data quality issues across enterprise systems?

Critical Realism (CR), as an ideal paradigm for technological phenomena (Mingers et al., 2013), will be used as the research paradigm and Affordances, as a practical guide to identify generative mechanisms using CR (Bygstad et al., 2016), as a lens to identify these generative mechanisms in their duality of human agents and the structures that they operate within over a fixed period of time. The six principles for using Affordance Theory in Information systems (Volkoff & Strong, 2017) would be used to analyse data collected from the case study. An explanatory approach to investigating the generative mechanisms causing poor data quality of master data would be used to ascertain why they came into being. An explanatory approach aligns with CR (Parra, 2023).

The phenomena would be generalised through empirical data from one case study of a Research Information System that was identified at a university in Africa to obtain both qualitative and quantitative data to identify and explain generative mechanisms in that situation. One case study was chosen as Type ET Generalizability indicates that adding more case studies will not enrich the generalizability of the phenomena under observation (Lee & Baskerville, 2003). The RIS implementation ran from 2016 to 2020.

Mixed-method research would be used to observe the underlying mechanisms involved in the generation of master data of poor data quality as this aligns with CR studies (Mingers et al., 2013). The data collection included 20 semi-structured interviews with key internal and external stakeholders as well as documentation in the form of reports, meeting minutes and presentations from the implementation team. Support calls would be analysed to include end user issues. Interviews were recorded and then transcribed, call logs extracted from the call logging system, and project documentation collected. NVivo would be used to assist with the coding.

Through the identification of the generative mechanisms behind the data quality of master data in an enterprise system, a deeper understanding of the possible causes of master data quality can be ascertained before organisations embark on future enterprise system implementations.

ABSTRACT. With the growing need for humanitarian assistance (Urquhart et al., 2022) and the domination of humanitarian funds and operations by international non-governmental organizations (INGOs) (Robillard et al., 2021), the World Humanitarian Summit decided to make humanitarian operations as local as possible as international as necessary, by involving local aid agencies (Lucatello and Gómez, 2022) to increase the effectiveness and efficiency of response operations. However, the move towards localization came with issues shaping information management (IM) activities (Nespeca et al., 2020), such as corruption tied to unreliable survey data (Altay and Labonte, 2014), disregarding the humanitarian principles (Comes et al., 2020), poor data management practice (Madon and Schoemaker, 2019), and systematic data bias in Yemen (Paulus et al., 2023) by local partners and actors. Evidence from Yemen exemplifies the military groups influencing the choice of survey location, survey designs, change of information on beneficiaries’ list, the gender-related and cultural taboos resulted in eliminating survey variables, or selecting beneficiaries based on certain social dependencies; the above issues show that elements of power and culture can contribute to shaping IM activities. This research uses activity theory (AT) [Fig. 1] as the lens to look into the possible issues, or here referred to as the “contradictions” for theorizing the existing challenges (Engeström and Sannino, 2011) by capturing structural tensions emerging across activities/actions that manifest themselves as the problem, ruptures, breakdowns, clash (Kuutti, 2019), conflicts, critical conflict, double-blind, and dilemmas (Engeström and Sannino, 2011), where its emergence leads to deviation from the established outcome of the activity; saving lives and alleviate suffering (OCHA, 2023). AT also provides a lens to locate social roles [communities], cultural norms & values [rules and regulations], and social structures [division of labour], which can, in turn, contribute to explaining contradictions. With the evidence of power and cultural norms and values behind the issues, ending up in aid diversion, corruption, marginalization from aid, and importantly collection of biased information, this study aims “To identify the contradictions and explain their causes concerning power along with social structure and cultural norms/values across IM activities within the localized humanitarian response operation in Afghanistan”. The research chooses Afghanistan due to its sociocultural complexities with the existence of local power brokers and community leaders influencing the humanitarian response (Zürcher, 2012), strict gender-based values shaping the aid delivery, biased information collection influenced by local power structures (Rietjens et al., 2007), and the role of Taliban in favouring their fighters and associated groups (Khosrow et al., 2023). The importance of this research lies in the phenomenon of localising humanitarian response, gaining interest due to global drivers such as the grand bargain (Frennesson et al., 2022), highlighting the need for empirical evidence to guide the humanitarian sector. Additionally, its rich context, localized humanitarian response operation in Afghanistan, involving various organizations and social roles [elders, religious Imams, gov. officials, people in need, and others] can help this research in theorising the relationship between power, social structure & cultural norms/values, and contradictions. A qualitative interview is used, following an explanatory model of social science [Fig. 2] (Danermark et al., 2019) rooted in critical realism as the research framework to identify and explain the contradiction across the IM activities. To further provide depth to the understanding of cultural norms & values and social structure, Archer’s morphogenetic approach [Fig. 3] (Archer, 1995; Archer, 2012; Archer, 2014) is employed as the social theory positioned within the explanatory model (Danermark et al., 2019) to look into the process of reproduction and change of social structures and cultural norms & values. Theoretically, this research follows the debates on the systemic vs periodic power perspective (Lawrence et al., 2012), and debates around the manifestation of power, their interplay, and understanding its effects in information system studies (Simeonova et al., 2020), supplemented with the literature on locating power across activities by employees in organizational context (Kelly, 2018), and the systemic view of power being related with episodic manifestation (Simeonova et al., 2024). Current research steps into depth by looking into the power of cultural norms and values attached to social communities/structures using a systemic view of power and understanding their influence on IM activities to explain the contradictions. This research also adds to the current literature by not only identifying the social structure and cultural norms & values but also providing insight into their reproduction and change, possibly leading to the emergence of new social structure and cultural values & norms across IM practices, with an emphasis on the role of agential reflexivity (Archer, 2007). Methodologically, this research extends the use of activity theory (Allen et al., 2013) by integrating the morphogenetic approach (Archer, 1995) to explain the change and reproduction of social structures and cultural norms and values. So far, 19 in-depth interviews have been held with employees of both national and international NGOs in Afghanistan, totalling 26 hours in length. The interviews are transcribed and coded, unpacking manifested contradictions such as refusing to respond survey questions, exaggeration in response, refusing to enlist people in the primary beneficiaries’ list, lying, favouring certain people and locations for survey, and disallowing women to participate in surveys to name a few. The key findings reveal the emergence of cultural values and norms namely Adaptive opportunistic response, aid entitlement culture, equal aid redistribution culture, favouritism, and corrupt information practices as the key cultural norms that emerged, explaining contradictions across IM activities. The findings also refer to the reproduction of cultural norms and values such as give-and-take culture, nepotism, irresponsible volunteerism, project completion over quality, patriarchy, and honour and modesty towards women. Additionally, shadow corrupt networks also emerged alongside existing social structures to name a few ethnicity, language, tribe, gender, and religion.

ABSTRACT. A common cybercrime, phishing uses the human vulnerabilities to steal data and can be catastrophic for both individual and organisational cybersecurity. Students as frequent digital platform users are a highly vulnerable group especially in the culturally, technologically and educationally influenced places such as Ghana. Anti-phishing solutions are improving, but hardly take into account context of the user and don’t really work for many diverse users especially in developing countries like Ghana.It is the objective of this research to study the factors that influence phishing vulnerability in Ghanaian university students and how we could build in the use of user context in machine learning algorithms to detect phishing. User context refers to demographic, behavioural and contextual factors like digital literacy, device use, culture etc. The project will analyse their impact on phishing vulnerability and build a contextually accurate, machine learning model to detect and prevent phishing attacks.The study is a pragmatist-led mixed-methods research with quantitative and qualitative methods. It will be an explanatory sequential design where we will begin with a survey asking students about awareness, experience, and experience of phishing. Selected random sampling will be used to achieve a more balanced sample and semi-structured interviews will offer deep qualitative findings. We’ll have the user logs and context variables as data to collect. Unsupervised and ensemble-based ML models will be trained on this data and evaluated against it for phishing detection. Principal algorithms such as Support Vector Machines (SVM), Random Forests, Neural Networks will be simulated and tested to use user context to reduce false positives.Qualitative phase will address students’ subjective experiences, culture, and technology1influences on their reaction to phishing. Combine this with survey quantitative data and results can be triangulated, giving both statistical and contextual information. Context added to machine learning not only helps understand the user context better – it helps create a personalized detection system.In theory, this study will add value to cybersecurity literature by demonstrating user context during phishing detection, building on models such as Protection Motivation Theory (PMT) to learn about users’ behaviour in cyber-threat scenarios. PMT is the means to analyse threat ratings and coping processes as they relate to users’ anti-phishing behaviour.Methodologically, this study uses a new context-informed machine learning model to fill the missing information on personal phishing detection and integrate behavioural, demographic,and environmental variables. By narrowing down to a certain culture and tech, it enriches international cybersecurity research.In real terms, the research can be applied to create smart anti-phishing products that will be responsive to users based on the user group, improving the online security of students and others. We can use the findings to create targeted awareness programs and cybersecurity education programmes for policymakers and educators. The results will help lower risk rates for university students, and to learn best practices for user context analysis of cybersecurity products.This study has also implications for future ML architectures that consider technical efficiency and UX. It places an emphasis on easy to use interfaces and culturally appropriate interventions, which fills a gap in existing cybersecurity tools. This will help to fuel the creation of novel adaptive cybersecurity solutions to make digital engagement more secure and trusted for weaker populations.This paper fuses theory and practice for a holistic view on phishing vulnerability. Bringing multidimensional user context into machine learning frameworks creates a baseline for future research on personalised cybersecurity interventions. The results are likely to be not just in terms of phishing detection, but also with regards to experience and trust in digital platforms. The paper also provides the empirical proof of the utility of ML-US variables in a multi-disciplinary context. These findings are of use to researchers, businesses and policy makers as part of a full-spectrum strategy to defeat phishing.