Embodied AI applied to Education

What is it about?

This paper, co-authored with Wendy Ross (London Metropolitan University) and Simon Penny (University of California, Irvine) and published in ACM Transactions on Computing Education, critically examines the intersection of Artificial Intelligence, Education, and Cognitive Science in ‘If you can’t dance your program, you can’t write it’: Challenges and Implications for AI in Education (title inspired by Tom Jennings). We argue that contemporary education risks becoming an automated repository of responses generated by systems such as ChatGPT, while sidelining the body, emotion, and lived experience. The consequence is a growing “cognitive debt” and a loss of epistemic agency. In response, we advance the concept of Embodied AI as a radical alternative. Drawing on Turing’s distinction between symbolic and sensorimotor approaches to machine learning, we highlight how current AI remains confined to the symbolic paradigm. Our analysis reframes AI in education through the lens of 4E Cognition, critiques the “omniscient oracle” model, and rejects a conception of creativity reduced to rapid outputs. We also address the “sensorimotor debilities” of digital technologies and revisit Arendt’s notion of the “banality of evil” to describe the risks of educators and students delegating critical judgment to opaque algorithms, thereby undermining education’s emancipatory potential. We conclude by proposing seven design principles for the critical and transformative integration of Embodied AI into educational practice.

Featured Image

Photo by Nubelson Fernandes on Unsplash

Photo by Nubelson Fernandes on Unsplash

Why is it important?

Reframe the Human–Computer Interaction Framework of GenAI from the Perspective of Embodied AI

Perspectives