Rethinking AI Through Embodied Art Practice

What if AI was not just a tool or a set of data-driven processes, but a co-created experience shared between humans and machines? This question drives “Somabotics: Creatively Embodying AI”, a project bringing together artists, researchers and technologists to challenge conventional assumptions about AI, robotics and human meaning-making. 

Led by Professor Steve Benford at the University of Nottingham, this UKRI Turing AI Fellowship aims to enhance artificial intelligence to enable humans to "make meaning". It explores how people make meaning through art and embodied experience, partnering with award-winning artists to create a series of robotic artworks, from robots that embrace and groom humans, to ones that dance and play music with them. By investigating these systems using an artistic lens the Fellowship aims to create AI artworks that embrace ambiguity, evoke interpretation, and champion improvisation. Touring artworks supported through the Fellowship will inspire the creative industries with new forms of cultural experience, while engaging the public to reflect on the future role of AI in society, especially on how it might become more inclusive.

Through Somabotics, we investigate how AI can help preserve art that moves, performs, and interacts – forms that often defy traditional archiving methods. By combining artistic inquiry with technological innovation, we seek ways to record the participatory and embodied qualities of these works so their vitality is sustained for generations to come.

At the heart of the project is the idea that meaning emerges from embodied experience – not only from cognition, but from interaction, emotion and movement. By using somatic and co-design methods rooted in Human-Computer Interaction (HCI), the project aims to humanise AI and explore how humans can make sense of the world through physical, artistic encounters with machines.

A New Model for Media Art Documentation and Preservation

Documentation is a vital practice in the archiving and conservation of process-driven, interactive, performative, and hybrid artworks. It is both past- and future-facing: preserving a work’s history while offering insights for its future activation. Through this dual role, documentation adds significant value to artworks – particularly those generated by AI, where it can also serve as a training dataset in addition to supporting presentation, dissemination, and conservation.

This raises an important question: can AI help preserve art that evolves, performs, and interacts? As media artworks become increasingly dynamic and participatory, traditional documentation methods often fall short. Somabotics explores how AI and robotics can meet this challenge, developing new ways to capture the shifting, performative, and participatory nature of these works so that their essence endures alongside their material form.

During Transformation Digital Art 2025, Steve Benford, artists Lancel/Maat, researcher and filmmaker Richard Ramchurn, and professor Gabriella Giannachi participated in the panel “Art, AI, and Robotics: Documentation Challenges and Possibilities of AI for Media Art and Performance”. Together, they explored how artists are embracing AI to push the limits of documentation in order to reflect on the liveness and unpredictability of media art.

Project Aims

Somabotics builds on the practice of soma design, a methodology that centres the body in the creation of technology. Developed within HCI, this approach invites designers to feel their way through the possibilities of new technologies – literally sensing their implications through movement, breath and interaction.

Through its multi-disciplinary approach, the project seeks to:

• Collaborate with internationally renowned artists in an iterative, artist-led design process involving both humans and robots, leading to the development of new concepts, technologies, and touring artworks.
• Open space for exchange with the AI research community through residencies, symposia, international visits, a summer school, and the release of a data-driven soma bits toolkit.
• Engage wider publics, media, and industry partners through touring exhibitions, a media campaign, and participation in science festivals.
• Position embodiment as a key concern in AI research, advocating for approaches grounded in human bodily experience, and fostering critical, creative, and inclusive perspectives on emerging technologies.

Shaping Research and Public Imagination Through Art

The project involves ongoing collaboration with:

University of Nottingham, University of Exeter, artists Lancel/Maat, Blast Theory, Sarah Wookey, LI-MA, and more.