Aim

The goal of this studio is to explore and discuss the potential benefits and the open challenges of giving a physical form to Artificial Intelligence (AI), towards the definition of tangible, or graspable AI. In this studio, we want to carry out a hands-on exploration of the potential of tangible interaction to help people grasp how transformer-based AI models, such as Large Language Models (LLMs), make predictions. Starting from a basic understanding of the core mechanisms of these models – input/output embedding, positional encoding, self-attention, and multi-head attention, we will explore how tangible interaction properties, such as data physicalization and tangible manipulation, may help designing graspable interfaces for facilitating the understanding and control of LLMs. Fostering AI transparency and human agency, we believe that tangible interaction may help increase the trustworthiness of Large Language Models. Through brainstorming and prototyping activities, we aim at identifying effective tangible representations and manipulation techniques that will help users grasp and trust LLMs.  

Venue

This studio is held at the TEI'25 conference in Bordeaux: https://tei.acm.org/2025/ 

Proposed Schedule

9:00 AM - 9:10 AM: Icebreaker

• Brief tactile, auditory, visual, or embodied immersion exercise to heighten sensory awareness and set the tone for the studio.

9:10 AM - 9:20 AM: Studio Presentation

• Explanation of studio goals, structure, and expected outcomes.

9:20 AM - 9:30 AM: Participant Introductions

• Each participant briefly introduces themselves (name, affiliation, research area, and motivations for participating in the studio).

9:30 AM - 10:00 AM: Provocation Talk

• A brief talk on the foundational mechanisms of large language models (LLMs) and limitations of current explainability techniques.

• Emphasis on how traditional interfaces (screen, keyboard, mouse) overlook potentials for tangible, embedded, and embodied interaction.

10:00 AM - 10:30 AM: Visualization Tools Exploration

• Participants explore existing LLM visualization tools such as BertViz, Bbycroft, and LatentScope.

• Interactive walkthrough by organizers provides hands-on experience and insights into the tools’ capabilities.

10:30 AM - 10:45 AM: Break

10:45 AM - 12:00 PM: Creativity Session

• Participants form groups to brainstorm potential tangible prototypes to make LLMs more interpretable and explainable through multi-sensory experiences.

• Each group receives the same card set but focuses on different materials, drawing inspiration from a range of provided resources.

12:00 PM - 1:00 PM: Lunch Break

1:00 PM - 3:00 PM: Prototyping Session

• Emphasis on iteration and testing. Participants think critically about how multi-sensory experiences can enhance AI explainability, incorporating feedback from the group.

• Potential avenues include:

  • Multi-Sensory Feedback: Combining tactile, auditory, and visual outputs into a single prototype.

  • Narrative and Embodiment: Using storytelling or bodily engagement with materials to clarify AI processes.

3:00 PM - 3:30 PM: Break

3:30 PM - 4:30 PM: Group Pitch & User Testing

• Each group presents their prototype, demonstrating how tangible, multi-sensory design can offer new perspectives on AI explainability.

• Key questions:

  • What AI process or model does your prototype aim to explain?

  • How do the chosen sensory modalities or materials enhance understanding?

  • What challenges and opportunities emerged from the design process?

• Peer feedback session: Participants offer constructive feedback to each presenting group.

4:00 PM - 4:45 PM: Group Reflection & Future Direction

• Discussion of insights and takeaways from the studio, focusing on material-based interactions and embodied sense-making in AI.

• Consideration of future research directions: How can these insights be integrated into ongoing XAI and HCI studies?

• Discussion of potential follow-up collaborations, publications, or future workshops.

4:45 PM - 5:00 PM: Closing Remarks & Next Steps

• Wrap-up of the day’s work, highlighting key points and next steps.

• Outline plans for documentation, sharing insights, and potential future projects.

Organizers

Leonardo Angelini is Professor at the University of Applied Sciences of Western Switzerland in Fribourg (HES-SO), teaching machine learning and human-computer interaction at the School of Management and the School of Engineering and Architecture. He ran several workshops in interaction design, including workshops on tangible interaction with the Internet of Things at Ubicomp’16, ETIS’16 and ETIS’17, TEI’18 and CHI’18, workshops on full-body and multisensory interaction at TEI’16 and Ubicomp’15 and ’16, and a workshop on tangible XAI at ETIS’22.

Goda Klumbytė is a postdoctoral researcher working on intersections between feminist epistemologies, machine learning systems design, and human-computer interaction. She works with transdisciplinary methods that foster critical technical practice in AI, leveraging critical perspectives towards social justice. She is currently working on feminist perspectives towards explainable AI.

Maxime Daniel is an Assistant Professor at ESTIA Institute of Technology in southwest France. He earned his PhD in Computer Science in 2018, contributing to the field of Human-Computer Interaction. His research focuses on the feasibility, utility, and usability of tangible interfaces with shape-changing capabilities, aiming to enhance their relevance and impact in human society.

Nadine Couture is Full Professor at ESTIA Institute of Technology in southwest France, where she also serves as Director of ESTIA-Recherche, the institute's research lab. She earned her Ph.D. in Enumerative Combinatorics in 1994, and her Habilitation à Diriger des Recherches (Tenure) in Human-Computer Interaction in 2010. She teaches core computer science courses to engineering students, including algorithms, web technologies, programming languages, mobile systems, augmented reality, and human-computer interaction (HCI) for aeronautics. Since 2001, her research has focused on Tangible Interaction, exploring areas such as the physical embodiment of data and whole-body interaction. In 2014, she founded the European Tangible Interaction Studio (ETIS) and has been a member of its steering committee ever since. In 2010, she also co-founded the PEPSS platform, dedicated to Human Factors in Interactive Technology. Nadine Couture actively fosters research-industry collaboration and serves as Vice President of CATIE (Aquitaine Centre for Information Technologies and Electronics).

Beat Wolf is Professor at the University of Applied Sciences of Western Switzerland in Fribourg (HES-SO) and part of the iCoSys Institute. Beat holds a PhD in Computer Sciences from the University of Würzburg. He has extensive expertise in machine learning and artificial intelligence (AI), applying these technologies across various fields such as handwriting recognition, oenology, circular economy, and genetics. Beat is co-founder of 2 start-ups: VCG.AI, an AI & big data platform accelerating the circular transition, and LYSR, AI-based anomaly detection for industrial processes monitoring and predictive maintenance.

Elena Mugellini is Professor at the University of Applied Sciences of Western Switzerland in Fribourg (HES-SO). She holds a Diploma (Bsc and Msc) in Telecommunications Engineering and a Ph.D. in Computer Sciences from University of Florence, Italy. Elena is the leader of the HumanTech Institute (Technology for Human well-being, humantech.eia-fr.ch/). Elena is part of the steering committee of ETIS - European Tangible Interaction Studio. Her research expertise lies in Human Computer Interaction and Human-centered Artificial Intelligence. Elena has been exploring natural and tangible interaction to support and facilitate human-AI collaboration and organizing various interdisciplinary workshops on Tangible Interaction and explainable AI (e.g. ETIS’22, CHI’21, AutomotiveUI’21, ICMI’20, AutomotiveUI’19, TEI’18, ETIS’17).

Claude Draude is a professor at the Faculty of Electrical Engineering/Computer Science, adjunct professor at the Faculty of Social Sciences, director at the Research Center for Information Systems Design (ITeG) and head of the interdisciplinary Participatory IT Design (PIT) research group at the University of Kassel. Her research includes human-computer interaction, socioinformatics and sociotechnical systems design and is characterized by her multidisciplinary background in cultural studies, media studies, design and artistic research, and informatics. She is interested in the way meaning and sense-making interconnect with the shaping of digital systems and how this relates to matters of participation, in- and exclusion, fairness and social justice.