https://youtu.be/6bfCuxI_84U?si=SbkwUWHfGVadypjI
https://arxiv.org/pdf/2605.24238
Toward Enactive Artificial Intelligence
Banafsheh Rafiee, Richard S. Sutton (2026-5-22)
Abstract
In this paper, we advocate for incorporating enactive approaches to perception and cog-
nition into artificial intelligence (AI). Enactive approaches view perception as an active,
skillful engagement with the world, where agents perceive by acting and by understand-
ing how their actions shape their experience. This contrasts with classical views that
treat perception as a passive internal process in which the brain receives sensory input,
processes it, and issues commands for action. Enactive views emphasize the dynamic,
embodied, and interactive character of perception, grounded in the lived experience of
agents embedded in their environments. We identify and develop four key enactive
concepts that we find most relevant to AI: experience, action–perception inseparability,
autonomy, and embodiment. Much of mainstream AI, from classical rule-based sys-
tems to large language models, has largely neglected these insights, treating cognition
as internal processing detached from embodied interaction and intrinsic normativity.
Reinforcement learning (RL), however, exhibits structural resonance with enactive prin-
ciples through its emphasis on action, agent–environment interaction, feedback-driven
adaptation, and agent-centered evaluation. However, this resonance should not be taken
as theoretical equivalence, as RL approximates some enactive insights,but key elements
remain absent or weakly developed. Building on this analysis, we suggest a broader
incorporation of enactive ideas into both mainstream AI and RL.
Enactive cognition is a theory in cognitive science proposing that the mind is not an internal computer that passively processes an objective reality. Instead, cognition arises through dynamic bodily actions, where organisms actively "enact" and shape their environment through continuous sensorimotor engagement. [1, 2, 3]
- Embodiment: The mind is deeply connected to the physical body. Brains do not exist in isolation; they are tools that expand an organism's capacity to navigate the world. [1, 2, 3]
- Action-Specific Perception: We perceive the world not by making an internal map of it, but by understanding what we can do with our surroundings (affordances). [1, 2, 3, 4]
- Sense-Making: Rather than just consuming information, living beings actively create "meaning" based on their needs, goals, and biological autonomy. [1, 2, 3]
- Participatory Sense-Making: Social interactions are fundamentally enactive. Our brains and minds co-regulate and co-develop through face-to-face engagements and shared attention. [1, 2]
- Sensorimotor Enactivism: Focuses on analyzing perceptual consciousness through bodily interaction, explaining how our movement allows us to "see" and experience the world. [1]
- Autopoietic Enactivism: Emphasizes the deep continuity between life and mind, focusing on how a living system self-maintains and self-organizes. [1, 2]
- Radical Enactivism: Rejects the idea of mental representations altogether, favoring an explanatory approach built on dynamic, embodied interactions rather than stored internal symbols. [1, 2]
- Foundational Readings: You can explore the Wikipedia Enactivism Page for historical context, or dive into overviews of Embodied and Enactive Approaches to Cognition on Cambridge Core.
- Expert Discussions: To see enactivism debated in real-time within the scientific community, you can watch A Masterclass on Enactivism and Cognition on YouTube featuring leading scholar Shaun Gallagher. [1]