View: session overviewtalk overview
| 09:00 | Generative Cultural Narratives: A Two-Stage KG-Guided Pipeline for Culturally Credible IDN PRESENTER: Angeliki Chrysanthi ABSTRACT. Generative storytelling systems increasingly rely on Knowledge Graphs (KG) to hold narrative state, yet the graphs they use are typically extracted on the fly and carry no provenance, which makes them unsuitable for cultural heritage contexts where stories are grounded on documented sources. We present a two-stage pipeline that combines event-centric semantic modeling with knowledge-graph-guided generative storytelling. We test this methodology on a demanding case study, the Monkey King (Sun Wukong), a figure who is worshiped as a deity, disciplined as a Buddhist pilgrim, celebrated as a literary rebel, and re-imagined across contemporary media, playing a different role in each of these cultures. This paper presents the methodology through which the tension between generative fluency and cultural fidelity is reduced, and the very act of representing a single figure across cultural variations is turned into an interactive and traceable mechanic. The paper also presents the promising preliminary results of the pilot evaluation carried out in two conditions that provides initial support for the method as well as a critical reflection of the suggested methodology. |
| 09:30 | Historical Empathy through Vulnerability: an IDN on the Great Siege of Malta (1565) PRESENTER: Jonathan Barbara ABSTRACT. History is written by the victors, and the factions on all sides of a con-flict are usually flattened into heroes or villains. This paper explores whether an Interactive Digital Narrative (IDN) can instead expose the vulnerability shared by all parties to a historical conflict. The IDN design removed power from the play-er, replacing it with helplessness and irreversible choices. The interactor plays a Maltese boy searching for his lost cat across three phases of the siege and fall of Fort St. Elmo during the Great Siege of Malta in 1565: inside Fort St. Elmo with the Knights of St. John, inside the Ottoman camp, and finally through the ruins of his home village. A four-group Solomon-style experiment (n=20) compared the IDN against a static text of the same narrative, producing higher empathetic scores across 16 of 17 questions, reaching significance for the Knights of St. John. Vulnerability was recognised unevenly: mostly towards the Knights and the Maltese civilians, and least towards the Ottoman soldiers, though interaction with the Ottoman NPCs reduced hostility. Within this exploratory sample, the findings suggest that interactivity appears to help players recognise the vulnera-bility of historical actors more effectively than a static text, though these findings should be treated with caution given the small sample size (n=20). |
| 10:00 | Reclaiming Empathy PRESENTER: Mattia Bellini ABSTRACT. Virtual and Extended Reality been popularly framed as the "ultimate empathy machine", a formulation that has proven remarkably durable in industry and public discourse despite sustained scholarly critique. This paper argues that this framing rests on a conceptual confusion: it collapses empathy into something closer to emotional contagion or projected identification. Tracing empathy's genealogy from the German aesthetic tradition through its extension to social sciences and history, we show that empathy was never conceived as an experience of merging with another but as a precarious, self-aware act of projection that presupposes and preserves the irreducible distance between self and other. We examine how VR and XR marketing and popular discourse eroded this structure, and how scholars in media studies, journalism ethics, and Interactive Digital Narrative research have identified the risks that result from the marketable distortion of empathy. Finally, we argue that recovering empathy's original philosophical structure, rather than abandoning the concept, offers a more productive theoretical foundation for designing VR and XR experiences intended to represent the condition of others, particularly in memorial and historically sensitive contexts, where the concept of historical empathy already operationalizes this structure for pedagogical use. |
| 10:15 | Representing the Past in Mixed Reality: Mimesis, Indexicality, and Metaphoricity for Memorial Works PRESENTER: Mattia Bellini ABSTRACT. The transmission of historical memory poses a fundamental challenge in the contemporary era: how can we engage with a past that no longer has living witnesses, whose material traces are fading, and whose complexity resists simple narrativization? This challenge is especially visible in cases such as Nazi internment and transit camps of the Second World War, where the ethical stakes of representation are high and the risk of trivializing immense human suffering is ever present. This paper presents a reflection on the representational affordances of Mixed Reality (MR) Interactive Digital Narratives (IDNs). More specifically, we propose a typology of representational approaches that can be adopted in IDN design for Memorial Work to foster contextual, perspectival, and affective comprehension. We discuss how these representational approaches apply to MR IDNs focused on representing the Nazi crimes of the Second World War, which benefits each approach provides, and which drawbacks each might present. Within each approach we present a prototype scenario developed by the project consortium for the same historical event, simultaneously exemplifying the framework's employment and reflecting on its use. The article moves from and is inserted within the framework of an experimental project aimed at exploring MR IDNs to represent life in [a Nazi camp during the Second World War - name redacted for peer review]. |
| 11:00 | Echoing Incentives: “Collaborative Interactive Design” and the multi-modal story-experience Sea Cipher ABSTRACT. A new research area is emerging, one that sits at the intersection of distributed authorship and participatory theatre. We call this ‘Collaborative Interactive Design’ or ‘CID’. It involves creating experiences that emerge at the nexus of two axes of interaction: planning and performance. Although the outcomes produced by this process are mercurial and evanescent, they often leave an indelible impression on creators and participants due to their resonance with key motivational factors. The authors were part of a creative team producing the multi-modal story-experience Sea Cipher. This dual-authored practice-as-research account has been written from two distinct epistemic positions: project lead and co-creator-actor. It draws on data captured in field notes and detailed production diaries informed by multiple motivational theories (Ryan & Deci, 2000; Schuler & Namioka, 1993; Dourish 2001; Katz, Blumler & Gurevitch, 1973). These are tied to the collaborative planning and interactive performance of Sea Cipher. Planning and performance are viewed as mutually influential and equally complex. Three ‘Incentive Dyads’ are identified: Aligned, Opposed and Mixed. The study considers how these pairings can be deployed to generate specific modes of interaction, thereby influencing the structure, tone, themes and character dynamics of interactive stories. |
| 11:30 | Adapt/Abilities: Truth, Avoidance, and Corruption as Pattern-Based Narrative Agency in Between the Tides ABSTRACT. Between the Tides is a psychological horror interactive narrative developed as a postgraduate design research project to address how pattern-based narrative adaptation can support player reflection on culturally situated responses to uncertainty, grief, and truth. This paper adopts the same question as its analytical lens, presenting a design-theoretical analysis of the project's core adaptive mechanism, the Truth-Avoidance-Corruption (TAC) system, which was developed and implemented in this research. Rather than modelling player agency through isolated moral choices, TAC proposes to treat repeated player behaviour as a form of narrative identity, tracking recurring patterns of engagement with and withdrawal from difficult knowledge and expressing them through memory, perception, dialogue, corruption, and atmosphere. The system draws on existing literature to frame grief as an oscillation between confrontation and avoidance rather than a linear progression toward resolution, and on critiques of visible morality meters to justify a hidden, diegetically expressed variable structure. The paper argues that TAC's non-surfaced, diegetic variables offer one way to represent such culturally situated responses without imposing a universal moral hierarchy on avoidance and confrontation. The propositions herein developed are intended for empirical testing in future player studies. |
| 11:45 | Transgressive Role Reversal and Perceived Learning in an Anti-Fraud Interactive Narrative Game PRESENTER: Jiawei Dai ABSTRACT. Transgressive role reversal can make educational interactive narratives experientially vivid, but it also asks players to inhabit morally troubling forms of agency. This paper examines this tension through “Revenge on Gold Diggers”, a Chinese full-motion video interactive narrative in which players first experience romance-scam victimhood and later become a deceptive counter-agent who manipulates targets framed by the game as culpable “gold diggers.” Based on semi-structured interviews with 20 players and reflexive thematic analysis, we develop a conditional learning uptake model. Findings show that narrative reversal did not directly produce anti-fraud learning. Instead, it repositioned players into deceptive agency that had to be morally appraised and narratively made sense of. Players responded through three modes of transgressive engagement: integrated acceptance and immersive engagement, negotiated acceptance and selective immersion, and critical distancing or non-enactment. These modes shaped whether anti-fraud learning was affirmed as procedural recognition of manipulation tactics, limited by doubts about realism or moral legitimacy, or refused as an unstable educational claim. The study contributes to interactive digital storytelling research by theorizing transgressive role reversal as moral repositioning and by showing how perceived learning in socially critical interactive narratives is interpretive, conditional, and morally mediated. |
| 12:15 | Recording the Cut: An Archaeological Standard for Interactive Digital Narrative Traversals ABSTRACT. Interactive digital narrative preservation has developed methods for systems and products, but not for the traversal itself. This paper proposes a traversal recording standard that treats foreclosed possibilities as essential evidence. Adapted from archaeological single-context recording and the Harris matrix, the method records three unit types: states, nodes, and closures. States are witnessed content. Nodes are moments of standing possibility. Closures are removals: selections, expiries, endings, quits, patches, and shutdowns that close what might otherwise have occurred. The closure is the decisive import. By numbering absences, attaching foreclosure lists, separating observation from inference, and arranging units in strict succession, the standard makes a traversal reconstructable, disputable, and comparable. A worked example records a ten-second Twine work whose traversal is dominated not by what is seen, but by what is lost. A correlation method then aligns multiple traversals as trenches on one site. The result is not another account of play, but a record format for the unpreserved middle of interactive digital narrative. |
| 12:30 | Becoming a subject: The reconfiguration of the digital self in the video game Red Dead Redemption 2 ABSTRACT. This paper argues that story-driven, open-world video games are privileged sites for the study of contemporary subjectivity, and it advances a theoretical model of the player-subject as a narrative identity that is embodied, emergent, and technically mediated. With Rockstar Games' Red Dead Redemption 2 (2018) as a case study, the study asks how interactivity and procedural logic reconfigure the construction of narrative identity, and what this co-creative process reveals about processes of subjectivation in digital culture. The problem is located in a specific textual phenomenon: the game's honour system, which converts each player action into an act of self-representation and answerability, thereby staging in operational form a question that narrative theory has treated only conceptually. Drawing on Ricoeur's theory of narrative identity, Foucault's account of subjectivation and of the classical mathesis, Deleuze and Guattari's notion of becoming, and the framing of the Digital Humanities as a transdisciplinary space, the paper develops an expanded hermeneutics able to read simultaneously the semiotic and the procedural strata of the interactive text. The analysis shows that the identity of the avatar Arthur Morgan is neither given in advance nor fully authored, but is configured procedurally in the act of play, within a rule system that both constrains and enables meaning. The paper contributes a vocabulary for the socio-cultural and critical study of adaptable interactive narratives, and it connects this vocabulary to the conference theme of adapt/abilities by reframing adaptability as an ontological property of the player-system coupling rather than a mere design feature. |
| 14:00 | From Fiction to Self-Reflection: Co-Designing a Visual Novel for Emotional Regulation through AI-Supported Rapid Prototyping PRESENTER: Catia Prandi ABSTRACT. Digital interventions can complement psychotherapy by supporting the acquisition and consolidation of therapeutic skills beyond clinical sessions. This is particularly relevant for approaches such as Dialectical Behavior Therapy (DBT), in which skills training and continued practice play a central role in helping individuals recognize, tolerate, and regulate emotional experiences. Yet maintaining engagement with repeated exercises and transferring therapeutic concepts to everyday situations remain important design challenges. In this paper, we present Breath, a visual novel serious game co-designed to support emotional regulation skills in young adults, with particular attention to individuals with Attention-Deficit/Hyperactivity Disorder (ADHD). Breath was developed through an AI-supported rapid prototyping process combining generative tools with human-led narrative and interaction design. This approach enabled us to rapidly materialize the narrative concept into an experiential prototype that was exploited during an expert-informed co-design focus group session with six mental health professionals. We discuss the potential of AI-supported prototyping for enabling early stakeholder involvement in the design of visual novel serious games for emotionally sensitive contexts, and how visual novel serious games can create a bridge between fictional interactions and self-reflection in emotionally sensitive domains. Preliminary findings highlight the perceived value of continuous narrative framing and game-based representations of emotional skills, while revealing design tensions in transitioning between fictional perspective-taking and personal reflection. |
| 14:06 | Ideator: Exploring AI-Supported Interactive Digital Narrative Authoring PRESENTER: Chidalu Ojiako-Chijioke ABSTRACT. Interactive Digital Narratives (IDNs) enable the representation of complex, multiperspective experiences, but authoring them requires structural reasoning that can challenge novice authors. Although generative artificial intelligence (GenAI) offers new opportunities for creative support, little research has examined how AI can support protostory ideation in IDN authoring. This study investigates how an AI supported mixed-initiative system can assist novice authors during IDN ideation. Using a Research through Design (RtD) approach, we developed Ideator, an AI-supported conversational agent, and evaluated it through qualitative user studies with participants with little to no prior IDN authoring experience. Findings show that participants valued Ideator’s conversational support and learned about IDN structure through interaction. The findings further suggest that ideation extends beyond an initial stage and continues throughout the authoring process. Participants were also willing to share creative responsibility with AI when transparency and authorial control were maintained. |
| 14:12 | From Novels to Interactive Worlds: Design Requirements from NovelVerse and DC-HMAS PRESENTER: Lan Zeng ABSTRACT. We investigate what design requirements emerge when iteratively materializing a reusable system intended to support high-agency, free-text interaction in a novel-derived world. Using a designer-led Research through Design approach, we examine two successive prototypes: NovelVerse and the Document-Centric Hierarchical Multi-Agent System (DC-HMAS). Drawing on code, saved interactions, and retrospectively reconstructed design documentation, we identify two situated requirements: using source-derived representations as context for GM-mediated interaction rather than primarily for narrative continuation, and coordinating privileged information with a consistent player perspective. DC-HMAS partially materializes these requirements through specialized agents, dual scripts, and Host/GM orchestration. Together, the reconstructed trajectory, the two situated requirements, and their partial architectural materialization constitute the paper's contribution. A recorded action-reinterpretation case further motivates consequential mediation as a direction for future implementation and evaluation. This would involve recognizing intended actions, judging them within world rules, and carrying relevant outcomes forward. |
| 14:18 | The AI Player as Narrative Design Partner: Exploring AI Participation in Tabletop Role-Playing Games PRESENTER: Akrivi Katifori ABSTRACT. Large Language Models (LLMs) are increasingly being used in tabletop role-playing games (TTRPGs) as Game Masters, non-player characters, and narrative generation tools. Comparatively little attention has been paid to the role of AI as a player. This paper presents an exploratory case study of a one-on-one text-based TTRPG session in which an LLM participated as a player character under the guidance of an experienced human Game Master. Drawing on gameplay transcripts, a post-session GM interview, and AI-generated reflections, we examine both the AI’s role-playing performance and its impact on the creative process. The findings indicate that the AI was able to collaboratively construct a character, maintain a coherent role-playing identity throughout play, and retrospectively articulate a character-development arc. Beyond gameplay, the AI functioned as a narrative design partner, acting as a design probe, co-creative catalyst, source of narrative resources, and reflective practice partner for the GM. We argue that the value of AI players emerges not primarily through content generation, but through sustained participation in collaborative storytelling. The paper concludes with implications for the design of AI-supported role-playing and interactive narrative systems. |
| 14:24 | AAT - A Tool for Narrative System Builders facilitating AI-assisted IDN Design ABSTRACT. We present AAT, the first - to our knowledge –open-source, pub-licly downloadable IDN authoring tool that offers GenAI support for both generation and runtime functions. Crucially, AAT makes GenAI optional so that creators can explore AI-assisted IDN design without it becoming man-datory. AAT combines a branching metaphor with a visual-programming paradigm and a procedural state system, is multimedia-native (image, video, sound, 360° panorama, and mobile XR), provides characters with a Big-Five-based affect system, fixed and responsive layouts, multi-language support, and Git-based version control. We further reflect on building AAT in collaboration with a GenAI coding agent, describing the practice as ‘su-pervised coding’ and emphasizing the need for human oversight. Finally, we discuss the new paradigm of narrative system building and what it means for creators. |
| 14:30 | Enhancing Grade School Science Learning through Interactive Storytelling with a Chatbot ABSTRACT. Traditional teaching methods often struggle to capture young children’s interest, making it difficult to foster understanding of scientific concepts. Research suggests that storytelling improves comprehension and engagement, and conversational chatbots offer personalized and interactive learning environments. Despite these individual benefits, limited studies have explored their integration to support primary science education through engaging and curriculum-aligned experiences. In this paper, we present Spark+y, an interactive digital storytelling environment featuring an AI learning companion designed to teach Grade 3 science concepts using a three-phase scaffolded approach. Spark+y leverages a pre-trained LLM to dynamically generate narrative content aligned with the government-prescribed MATATAG curriculum. Preliminary usability testing with 20 Grade 3 students demonstrated the high potential of an LLM-driven narrative-anchored AI learning companion as a pedagogical tool to enhance curiosity, comprehension, and accessible learning materials in early science education. |
| 14:36 | How Much Influence Is Enough? Investigating Narrative Interactivity in Child–Robot Interaction ABSTRACT. Storytelling plays an important role in childhood learning and development by providing opportunities to understand events, relationships, and social situations. Social robots offer an interesting medium for narrative experiences because they can engage children in embodied, face-to-face interaction while allowing them to engage with an unfolding story. However, narrative interactivity can take different forms, varying in the degree of control children receive, the form and depth of their interaction, and the consequences their contributions have for the narrative. At the same time, increasing children’s influence raises the question of how control over the overarching narrative can be maintained. This late-breaking report presents a research agenda investigating how characteristics of narrative interactivity in child–robot interaction affect children’s psychological and cognitive outcomes while maintaining narrative control. Across three studies, the project examines these characteristics from complementary perspectives. Study 1 investigates different degrees of narrative control and provides preliminary evidence for effects on narrative self-efficacy and story recall. Building on these findings, Study 2 will examine differences in the form and depth of narrative interaction, drawing on the ICAP framework and Generative Learning, with a focus on deeper cognitive processing and understanding of narrative content. Study 3 will examine narrative consequences by investigating whether the extent to which children’s contributions affect the narrative influences psychological and cognitive outcomes. Together, the studies aim to provide a differentiated understanding of which characteristics of narrative interactivity matter for children’s psychological and cognitive outcomes and how children’s influence can be balanced with control over the overarching narrative. |
| 14:42 | Student Perceptions on Teaching & Learning Games & Interactive Digital Narrative with Generative AI PRESENTER: Lindsay Grace ABSTRACT. The past few years have seen the rapid rise of generative-AI tools, models, and platforms, many of which have specific application to interactive digital narrative (IDN) creation. The authors have been teaching games and interactive narrative courses at both introductory and advanced levels for several decades and have recently sought to explore these new tools and technologies in their courses with students (or at least discuss their potential benefits and draw-backs in media creation). This paper explores both the prior inclusion of technology and automated workflows in IDN courses, the more recent inclusion of generative AI solutions, and student perceptions, attitudes, and reactions to these modern platforms. This paper is presented as an informal, observational review of our experiences, and given the sensitive nature of the subject, all quotes and data are anonymized. This work is intended a precursor to a more formalized study of student perceptions and practices, and so the emphasis here is on an initial review from multiple educators identifying similarities and differences in their classrooms world-wide. |
| 14:48 | Visualizing Biodiversity Futures: From Children’s Drawings to Interactive Digital Narratives through LLM Co-creation PRESENTER: Catia Prandi ABSTRACT. Climate change and biodiversity loss are reshaping ecosystems worldwide, making it increasingly important to explore how children understand and imagine environmental futures. To this end, we conducted a workshop designed to support children’s participation in the creation of biodiversity future narratives. The workshop was carried out in the [anonymized] park, a socio-ecological environment undergoing significant environmental transformation. The workshop integrated environmental education and a two-stage drawing exercise in which children depicted both the current and future state of the local environment. The resulting drawings were then used as prompts for Large Language Models (LLMs) to support the co-creation of interactive digital flipbooks, transforming children’s visual ideas into narrative experiences that bring imagined futures to life. The study contributes a child-centered methodology that integrates participatory drawing about environmental futures, and LLM-supported narrative co-creation to enhance creativity and reflection on biodiversity futures. Furthermore, the work explores how children’s drawings can serve as narrative seeds for developing richer Interactive Digital Narratives (IDNs), representing a promising direction for future research. |
| 14:54 | Small Model, Hard Constraints: Confining an LLM to Classification and Phrasing in a Locally-Run Interactive Narrative ABSTRACT. Interactive narratives are beginning to delegate character dialogue to language models, but the few exchanges that carry the plot cannot be improvised. In our case it is a contested account of what happened between two characters at a party, which the design deliberately never resolves. We present SCRIM, a system in which the model classifies player input and phrases authored output, and decides nothing in between. The design principle the system moves toward is that the graver a narrative element, the less of it passes through the model at either end: the conditions releasing it rest on measured state rather than interpreted tone, and its wording is authored verbatim rather than generated. SCRIM runs on a 7B model locally, with no hosted service. Replacing that model with a 14B one left character coherence intact but changed how one player message was interpreted at the input boundary, closing a single narrative gate. Because every condition is authored data rather than code, the divergence was traceable to a single field and adjusted there, after which the gate reopened with the model's reading unchanged. Ablating the output guards and the retry loop left the narrative structure untouched (the same locks fired at the same turns, and every pass reached the same ending) while the served text degraded, with speaker labels leaking and characters repeating themselves verbatim. The load sits on the authored conditions; the guards defend the form, which is what a player meets first. |
| 15:00 | Far Letter: an AI-Driven Asynchronous Epistolary Narrative Game ABSTRACT. Many current AI-driven narrative games use real-time conversational dialogue as their main form of interaction. While some games have introduced waiting between conversations, asynchronous interaction is rarely used as the fundamental interaction structure. This study explores the player experience of epistolary asynchronous interaction with an AI-driven character in AI narrative games. We developed Far Letter, an AI-driven narrative game prototype in which players communicate with an AI character through letters. The epistolary interface and mechanics allow each exchange to include a period of separation between writing and receiving a response, making waiting an important part of the interaction rather than an occasional interruption. It also supports a slower and longer-form style of communication compared with chatbot dialogue. Through a planned qualitative user study, this study will analyze and conclude players’ experience in this game prototype, with particular attention to their thought about pace of narrative, letter writing and emotional engagement. |
| 15:06 | Generative AI- Supported Podcasting for Public-Facing Game Studies and Culture Hacking ABSTRACT. This paper examines Game Studies for Everyone, an experimental podcast utilizing AI to broaden access to video game and interactive narrative research. By employing AI-generated summaries, voice synthesis, and automated publishing, the project transforms dense scholarship into accessible audio episodes. We analyze the production process and explore ethical challenges in AI-mediated storytelling, particularly regarding authorship and voice. Ultimately, this project applies culture jamming principles to democratize academic knowledge, contributing to ongoing discussions on generative media's role in reimagining academic accessibility. |
