Frictions, overlaps, and productive tensions: Embodiment in Creative AISaturday 14 November 2026, 10:00 – 11:15
Soft robotic environments: grounding embodied AI in physical experience and design questions
- Valentina SoanaLecturer; PhD Researcher The Bartlett School of Architecture, UCL; Mechanical Engineering, UCL
In the development of soft robotic environments, embodied AI becomes a subject of design inquiry, grounded in structural behaviour, material intelligence and human experience. What if architecture were no longer predetermined, rigid and static, but soft, intelligent and capable of transforming in response to human and environmental conditions? How might such environments embrace the diversity and unpredictability of lived experiences, opening up new possibilities for interaction?
These questions are explored through the interdisciplinary development of Elastic Robotic Structures (ERS) for interactive environments, advancing a vision of soft robotic architecture. ERS are lightweight, shape-morphing systems integrating architectural design, structural mechanics, soft robotics and computation to create responsive, human-centred environments.
Within this approach, elasticity becomes both a structural and a design principle. Material behaviour, morphology and movement are intertwined with computational and control strategies, evolving through continuous negotiation between physical constraints, human interaction and architectural intentions. Beyond engineering applications driven by predefined tasks and performance metrics, designing soft robotic environments raises questions that are fundamentally conceptual, physical and experiential.
In this context, embodied AI offers the potential to extend sensing, control and adaptation, enabling structures to respond to changing conditions and human interactions. Yet the challenge is not simply what AI can generate or optimise, but what forms of human–material–machine relationships might emerge. Controlling large elastic deformations becomes an experiential design question: how might movement enhance architectural performance, accommodate diverse human needs and enable new forms of interaction?
Embodied AI is therefore neither an autonomous driver nor a predetermined solution, but a potential extension of an evolving design process. Its objectives, constraints and criteria for success emerge from architectural inquiry, physical experimentation and lived experience. What to create, how and why remain fundamental acts of design