Short Paper Showcase: Group B
Date: Monday 2 Nov 2026
Time: 15:10–15:25
Add this session to your calendar. Talks run back-to-back in the order below.
| Order | Presentation | Abstract | Authors | Contribution type |
|---|---|---|---|---|
| 1 | Disruptors and Co-Designed Interventions for Autistic and ADHD Users in the Audio-Visual Media Context | IP-based media distribution and machine learning enable increasingly personalised media, yet the accessibility needs of autistic and ADHD communities remain largely underserved. We conducted a multi-stage participatory design study (N = 8) with autistic and ADHD participants, combining interviews and co-design activities to identify media disruptors and develop targeted interventions. Through reflexive thematic analysis, we identified key disruptors across sensory, cognitive, and narrative modalities: impeded narrative comprehension, auditory discomfort, dialogue masking in multi-speaker scenes, aversive scene triggers, misaligned pacing, and transient text presentation. From 28 co-designed concepts, we synthesised six functional intervention clusters (spanning intersemiotic translation, audio alteration, content segmentation, and dynamic recaps) and evaluated interactive mockups with a subset of participants (N = 5). Evaluation revealed initial evidence of high perceived utility (Mdn = 3–5), while highlighting that participants required flexible control to adapt interventions to their personal preferences and viewing contexts. This work provides preliminary design insights for integrating user-tailored accessibility for autistic and ADHD audiences into streaming platforms, laying the groundwork for future generalisable evaluation. | Cieciura, Glancy, Jackson | Short paper |
| 2 | Towards AI Design Companion: A Three-Stage Framework for Reflective Design Learning with Generative AI | Generative artificial intelligence (AI) offers new opportunities for design education, but generic generative interfaces give novice learners powerful outputs without explicit support for how to explore, question, and revise them. This paper presents AI Design Companion, an early-stage web prototype for conceptual design activities in which learners expand an initial idea, interrogate assumptions and constraints, and refine a concept proposal. Drawing on reflective practice, creativity support, and recent HCI research on AI-assisted ideation, the prototype organises AI assistance into three iterative roles: a cognitive catalyst, a reflective mirror and sparring partner, and an assistant/evaluator. The paper contributes (1) a theoretically grounded human–AI interaction framework instantiated in a working prototype and (2) preliminary design-refinement insights from a structured four-model proxy critique conducted before human evaluation. The proxy exercise identified needs for stronger reflection prompts, clearer progress cues, and explicit controls for adopting, adapting, rejecting, or further investigating AI suggestions. The work argues for AI interfaces that scaffold novice design judgement while keeping learners responsible for design decisions. | Dong, Castagna, Suleimenova | Short paper |
| 3 | Designing ValoTheko: A Formative Needs Assessment and Low-Fidelity Design Concept Study of a Conversational User Interface for Bangla Multimodal Mental Health Survey | Accurate data is critical for research, and surveys are the link connecting researchers with public perception. However, conducting surveys is challenging in Low and Middle Income Countries (LMICs) such as Bangladesh due to limited digital experience, literacy issues, and discomfort in disclosing sensitive personal information. These factors exclude underrepresented populations and create biases, especially in mental health surveys, where stigma and privacy concerns affect participation in digital interventions. To mitigate these issues, we propose ValoTheko, a low-fidelity design concept for a Bangla multimodal Conversational User Interface (CUI) for mental health surveys. ValoTheko is examined through a two-fold approach comprising a formative needs assessment followed by a low-fidelity prototype concept evaluation. In the first phase, snowball sampling was used to recruit 137 participants (male = 72, female = 65, age = 18-65) for a needs-assessment survey. Based on the findings, we derived design requirements and developed a low-fidelity prototype. The second phase involved a formative prototype concept evaluation conducted with 17 newly recruited (random sampling) participants (male = 8, female = 9, age = 18-65), where participants reviewed a non-interactive screen-by-screen walkthrough and provided their anticipated reactions rather than interacting with a working system. The findings indicate that mixed-mode interaction represents a promising approach for sensitive mental health data collection, as it accommodates both expressive voice input and click-based response mechanisms. | Remal, Chowdhury, Islam, Rahman, Zaber, Amin | Short paper |
| 4 | Beyond Screens: Exploring Alternative Forms of Personal Data Visualisation Through Co-Creation | While sensor technologies have grown and become democratised, so have the collection of data relating to the individual. But the presentation of such data, if it exists, is typically made interpretable to all. In this paper, we present results from the first of a series of co-design workshops, during which we prompted participants to reflect and co-create novel alternative forms of data representations that felt more personal. We found that personal data covers many categories, with no simple rules for what should be private, semi-private or public: in some cases, sharing personal data was described as desirable. This co-creation workshop allowed individuals to explore physical design spaces and produce their own abstractions. This is the first step to understanding such personal metaphors, leading to further exploration of reflexivity and actionability for individuals that goes beyond traditional visualisation practices and media. | Le Bras, Georgiou, Winton, Padilla | Short paper |
| 5 | Useful, Not Human: A Feasibility Study of User Trust and Credibility in AI Interview Training | AI-powered interview simulators are increasingly deployed as scalable training tools, yet little is known about whether users perceive them as credible sources of developmental feedback. This paper presents a mixed-methods feasibility study (n = 14) evaluating user trust, credibility, and feedback acceptance in a voice-based AI interview simulator, combining validated psychometric instruments with qualitative open-text responses in a pre–post design. Feedback delivery showed preliminary signs of being a failure-sensitive component, with participants who experienced feedback failures tending to rate helpfulness lower even when the interaction itself was perceived positively, though this rests on a small, non-randomised subgroup. Distrust decreased following a single session (r = .68) whilst active trust did not rise to the same degree, supporting a two-factor treatment of trust. Users also adopted functional mental models (Examiner, Tool, Coach) rather than relational ones. These preliminary findings suggest the success of AI interview training may depend less on human-like realism and more on reliable, contextually grounded feedback. | Johnston, Hudelist, Hannay | Short paper |
| 6 | Sense-O-Plex: How Compositions of Cross-sensory Artefacts Shape Children’s Collective Emotional Sense-Making | Children increasingly encounter sensory-rich technologies in classrooms, museums, and play, yet little is known about how they collectively negotiate meanings across vision, touch, and smell while shared references and vocabularies are emerging. We introduce Sense-O-Plex, in which 52 children aged 8–11 worked in small groups to compose shared cross-sensory artefacts from visual, tactile, and olfactory items on an affect circumplex. Qualitative analysis of talk, gesture, action, and decisions shows that colours supported rapid consensus through familiar cultural associations; shapes prompted embodied exploration and negotiation; and scents exposed limits of shared reference, producing personal stories, disagreement, or conversational stalling. We contribute an account of how artefact composition, shareability, and manipulability shape participation, explanation, disagreement, and closure, with implications for collaborative sense-making technologies for children. | Li, Roberts-Morgan, Lo, Tierney, Horton, Read, Metatla | Short paper |
| 7 | Beyond Generation: How Creative Practitioners Use GenAI to Articulate, Evaluate, and Preserve Creative Judgement | Generative Artificial Intelligence (GenAI) is increasingly integrated into creative workflows, raising questions about its impact on human cognition and authorship. This exploratory qualitative study examines how experienced creative practitioners engage with GenAI across creative and educational contexts. Our findings suggest that participants did not experience GenAI as simply reducing or preserving critical thinking. Instead, they described critical judgement as occurring across the creative workflow: before generation, as practitioners articulate intent and translate tacit criteria into prompts; after generation, as they compare and interrogate outputs; and across iteration, as they verify uncertain material and decide whether outputs should be accepted, refined, rejected, or reframed. We identify five design opportunities for practitioner-led GenAI: intent articulation, prompt formulation, output comparison, verification, and boundary decisions. By shifting attention from automated production to reflective creative judgement, this work offers a design-oriented account of how participants used GenAI to support, rather than replace critical thinking. | Pham, Holmquist | Short paper |
| 8 | From Justice to Care: Envisioning Technology’s Role in Offline Misinformation Correction | The offline spread of misinformation remains largely overlooked in HCI research, despite a documented role in real-world harms. This paper situates the problem in urban environments, where existing technological infrastructures offer potential for large-scale intervention. Using a speculative co-design approach, we ran workshops asking participants to envision futures in which urban technologies detect and respond to the offline spread of false information. Using Activity Theory as an analytical lens, we identify two roles that participants envisioned technology taking in this context: mediating justice; and mediating care-giving. We discuss the additional challenges an offline, urban setting brings to false information tracking. | Scott, Mahoney, Colborn-Clark, Talhouk | Short paper |
| 9 | AI functional fixedness and the challenge of designing beyond efficiency | This paper examines what drives design students, even when explicitly invited to move beyond utility in AI artefact design, to drift back toward efficiency-driven outcomes. Drawing on two workshop series, it traces this recurring pattern across five annotated cases and identifies the cultural, cognitive, and material forces sustaining it. Building on the theory of functional fixedness, the authors extend the concept from the individual object to a structural, cultural level, naming the resulting condition AI functional fixedness. The paper closes by sketching directions for future pedagogical inquiry aimed at disrupting this condition. | Caruso, Cioni, Squatrito | Short paper |
| 10 | Ramadan Fasting Selectively Reshapes Cognition: A Webcam Eye-Tracking Study in the Wild | Ramadan fasting is observed by 1.8 billion people annually, yet HCI has not characterised how fasting reshapes cognition or what this means for interface design. We deployed a browser-based webcam cognitive assessment platform across 60 sessions from 18 participants during Ramadan 2026, capturing six oculomotor features and computing a four-dimensional Cognitive Readiness Index (CRI). Fasting is not associated with uniform suppression of cognition but with selective reshaping: Pattern Recognition significantly exceeded Visual Processing in 82.4% of high-reliability sessions (M = 47.3 vs M = 31.9; Wilcoxon W = 55, p < 0.001, n = 34). This divergence narrows beyond ten fasting hours (ρ = −0.452, p = 0.007), suggesting a practical design boundary for fasting-aware systems. A literature-anchored heuristic comparison against established oculomotor norms provides partial support for interpreting observed score levels relative to a non-fasting reference. Webcam capture quality must be explicitly stratified before inference (p < 0.001, r = −0.524). These preliminary findings have direct implications for the design of adaptive digital interfaces serving the 1.8 billion people who fast annually. | Tahri Sqalli, Meftah El Menani, Kanzi Belghiti | Short paper |
| 11 | Patterns of Soft Financial Delegation | Financial delegation involves entrusting someone to manage your finances and can be key for independent living for people with limited financial capability. Traditionally, the only legitimate way to delegate finances was through lasting powers of attorney (LPA). However, LPA fails to consider everyday needs of managing money. Recently, new delegation services are allowing people to flexibly delegate specific financial matters - a concept we call soft delegation. This paper reports on an analysis of nine soft delegation services, with the aim to understand the ways they enable soft delegation. Findings of our analysis demonstrate an understanding in the way the social challenges of financial oversight, configurability and power imbalances affect how we might use these services to delegate. | Khalilian, Elsden, Vines, Wilkinson | Short paper |
| 12 | To Stay or to Step Back? Relational Transitions in AI Companionship: Preliminary Insights from a Pilot Scoping Review | AI companions are conversational or embodied systems that users engage with in sustained relational roles, yet they are often evaluated through retention-orientated metrics that treat continued interaction as a marker of success. Human-AI relational involvement can change or end, and how such transitions should be understood and designed remains unclear. We report a pilot scoping synthesis of relational transitions in AI companionship, drawing on a preliminary corpus of 51 charted papers from an ongoing PRISMA-ScR review and iterative qualitative synthesis. Across the evidence, companion relationships evolve through distinct transition pathways, reflecting changes in user needs, relational value, safety considerations, and platform conditions. We develop a preliminary working framework for mapping relational presence pathways (not mutually exclusive): (1) appropriate continuation, (2) protective withdrawal, (3) supported graduation, (4) imposed discontinuation, (5) user-initiated exit, and (6) gradual fade. We further examine continuity work after transition, including user attempts to restore disrupted AI relationships and platform efforts to shape relational continuity. Current evidence clusters around user retention, risk management, and externally imposed changes, while adaptive and user-centred relational transitions remain insufficiently studied. Informed by psychotherapy termination literature and self-determination theory, we argue for a shift from retention-orientated design towards supporting safer relational transitions. | Xiao, Zhang, Hu, Calvo | Short paper |
| 13 | Raise, Don’t Replace: Two Kinds of Connection in AI Character Products | Some new consumer apps promise to do your socialising for you: an avatar that likes, comments, breaks the ice, and screens people on your behalf. We think this confuses two different things. Connecting with another person depends on that person actually showing up; hand the act to a stand-in and most of its value goes, which is part of why a message believed to be written by AI is trusted and felt less. Connecting with a virtual character is not like that. The character is openly made, and what carries the bond is what you put into it, not whether some hidden human meant it. Automating the first kind tends to hollow it out, and an avatar acting as you is unlikely to build the second kind either. The opening many products are missing, we argue, is that second kind: the bond with a character you make and raise. Drawing on a deployed AI-character game and 20 interviews with its creator-users, we describe what seems to carry that bond, authoring, a character that stays coherent, care over time, and seeing it in your own life, and report something we did not expect: most (though not all) of our most committed users were lukewarm on AI conversation, yet were deeply attached to their characters. | Zhu, Guo | Short paper |