This work in progress investigates human interaction with an LLM-powered chatbot, presented as either a fellow human or a transparently disclosed AI collaborator, in a high-stakes decision-making simulation—the NASA Moon Survival Task. We will employ a one-way between-subjects design to examine how individuals’ collaboration and communication are influenced by the identity of their partner (AI vs. human). Specifically, we will evaluate individuals’ collaboration processes (i.e., collaborative behaviour and communicative dynamics) and outcomes, alongside their retrospective interaction experience and perceptions of the partner. We will also examine dyadic-level linguistic coordination during the interaction and conduct user profiling to uncover variations in AI collaborative benefits. We anticipate that this study will have four key impacts: safeguarding human-AI collaboration, democratising AI benefits, guiding model improvement, and making methodological contributions. The anonymised dialogues and associated data will be open-sourced upon study completion.

<p xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" class="first" dir="auto" id="d931711e55">This paper presents the design and implementation of a nine-piece generative art installation, conceptualized as a family of intelligent shapes. Each shape is represented by a distinct 3D mesh that interprets and transforms a shared set of AI signals based on its unique morphology. The installation explores the hypothesis that ‘the way an entity is shaped affects the way it thinks’. An AI system broadcasts identical parameter signals to all meshes, but each shape processes these signals in a way that reflects its structure. The AI continuously monitors each shape’s deformation, creating a feedback loop in which morphological changes influence subsequent signal processing on the part of the AI. This results in a range of diverse evolutionary behaviours, demonstrating how form shapes computational ‘thought’ processes in autonomous systems. This interplay between uniform inputs and form-specific responses suggests that a geometric mesh’s physical properties can serve as inputs for reasoning, with emergent shape changes reflecting cognitive-like processes. The paper details the conceptual motivation, system architecture, technical implementation, and observed outcomes, offering insights into how morphology guides information processing in generative art and AI, and proposing new directions for form-driven intelligence.

<p xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" class="first" dir="auto" id="d317443e108">Given its sensory and emotional qualities, food as a rich resource for design has been much explored in Human-Food Interaction research. Besides such positive aspects of food and food experiences, other scholars have focused on problematic and disordered eating, as well as technologies’ negative impact on the growing number of people living with these conditions. Drawing from mindfulness-based interventions, health research has explored effective interventions for mindful eating, which however have been limitedly harnessed in interaction design, despite the growing body of HCI research on mindfulness technologies emphasizing also the role of the human body in mindfulness practices. This half-day workshop aims to address this gap, by bringing together designers, practitioners, and HCI researchers from human-food interaction, mindfulness, and somaesthetics to explore the design space of technologies for mindful eating interventions, with a focus on five goals targeting (i) support for the role of the body in mindful eating, (ii) design exploration of technologies supporting mindful eating interventions, (iii) develop theoretical and methodological foundations for designing for mindful eating, (iv) explore ethical aspects of designing for mindful eating interventions, and (v) position mindful eating technology-based interventions within a broader, more sustainable food cycle.

<p xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" class="first" dir="auto" id="d313717e70">Blind and low-vision (BLV) learners encounter significant barriers in accessing STEM content, as traditional accommodations, such as braille transcription and tactile graphics, are time-consuming and labor-intensive, often resulting in delays that impede timely engagement with STEM content. Responding to this accessibility gap, we present StemA11y, an AI-powered iOS application that automatically converts STEM materials into accessible formats and delivers them via speech, audio cues, haptics, and VoiceOver support. We evaluated StemA11y with 11 BLV participants across two content types–math worksheets and adapted SAT papers. Our results revealed that System Usability Scale (SUS) scores exceeded the industry benchmark of 68 for both math worksheets (M = 75.2) and SAT papers (M = 83.2). Technology Acceptance Model (TAM) ratings indicated high perceived ease of use, ease of learning, and overall satisfaction. Further, qualitative feedback demonstrated the importance of customizable content verbosity and flexible navigation. Drawing on these insights, we reconsider STEM accessibility for BLV learners and discuss design considerations for developing multisensory AI-powered mobile systems to improve STEM accessibility, including customizable content verbosity, strategies to reduce redundant audio cues, and enhanced support for flexible navigation.

<p xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" class="first" dir="auto" id="d4043734e66">This paper introduces a progressive speech-driven virtual reality (VR) interaction model that enhances user engagement and accessibility. Traditional VR interactions, often reliant on hand controllers, can limit user immersion and accessibility. Our proposed model utilizes partial speech recognition to incrementally unlock VR functions as users speak predefined or context-relevant sentences, contrasting with conventional single-command voice interactions. This approach is demonstrated in the immersive artwork "Unseen Claimed," where voice-based interactions replace hand controllers, allowing participants to engage with the system by speaking identity-related manifestos. These manifestos sequentially reveal and ultimately dissolve virtual representations of the self, encouraging participants to confront unfamiliar reflections and disappearing virtual personas. The artwork combines real-world acts of self-definition with conceptual investigations into notions of non-identity, promoting inclusivity and offering a reflective framework for examining the boundaries of selfhood within digitally mediated environments. Preliminary testing indicates that this speech-driven interface offers more intuitive interactions, reduces physical exertion, and enhances the sense of presence compared to traditional controller-based VR systems. The progressive speech recognition system provides continuous feedback, reducing cognitive load and fostering a deeper connection to the virtual experience. This research highlights the potential for voice-centred systems to enhance not only user experience and inclusivity but also conceptual explorations of virtual identity in VR contexts.

<p xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" class="first" dir="auto" id="d313249e77">This paper argues that user interfaces need to be explainable whether or not they contain artificial intelligence components. Even with the best design, complex applications often leave users confused; this is exacerbated on small touchscreens, where small slips can lead to markedly different outcomes and when notifications or intelligent agents may autonomously change the interface. This can be disorienting even for the most tech savvy user, but doubly so for those less confident or with motor-control issues. We are often left asking “what just happened?” or “how can I do this again?”. We need explainable user interfaces.

<p xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" class="first" dir="auto" id="d4045909e55">The UN describes current progress in closing the CO2 Emissions Reduction Gap as representing a “massive gap between rhetoric and reality”. To meet the goal of the Paris agreement “we would need global mobilization on a scale and pace never seen before. Many will say this is impossible.” I don't believe it's impossible but, to achieve the necessary turn-around, we have to unlock new and relatively unexplored digital opportunities for reaching net zero. What I'm proposing is a cybernetic approach to solving the problem. In 1959, the UK cyberneticist, Stafford Beer, prepared an analysis of comparative systems (i.e. what organizations do and how they do it) by grading them from simple to complex. He placed only three systems in his ‘exceedingly complex’ category: ‘the economy’, ‘the brain’ and ‘the Company’. Beer’s list was a starting point in defining each system's predictability and its susceptibility for control by tried and tested scientific methods. If he were to be alive today, I feel sure he would have included ‘Climate’ in his ‘exceedingly complex’ shortlist. This system, too, must be subjected to new theories of prediction together with proposed methods for testing the validity of these theories, in order to tackle the most formidable challenge of all time. I am advocating that research should be undertaken, as a matter of some urgency, into the idea that the digital attributes of the metaverse can help in lighting small fires of hope in an otherwise dire situation,