UCLA Study: AI Lacks Body Experience to Fully Understand Humans

Researchers propose 'internal embodiment' as key missing component in current AI systems

Apr. 2, 2026 at 2:04am

Got story updates? Submit your updates here. ›

A new study from UCLA Health argues that today's most advanced artificial intelligence systems lack the essential ingredients of a physical body and internal awareness that humans take for granted. The researchers propose that building functional analogues of 'internal embodiment' into AI represents a crucial and underexplored frontier in the field.

Why it matters

The paper states that the distinction between external and internal embodiment has significant implications for how AI systems behave and how safe and trustworthy they can become, especially as they are deployed in consequential settings. Without an internal regulatory mechanism, AI models can sound experiential without truly understanding human experiences.

The details

The study focuses on multimodal large language models like ChatGPT, which can process and generate text, images and video, but lack the bodily experience to truly understand concepts like fatigue, uncertainty or physiological need. This limitation was illustrated when several AI models failed to properly identify a simple point-light display of a human figure in motion, which even newborns can recognize.

  • The study was published in the journal Neuron on April 2, 2026.

The players

Akila Kadambi

A postdoctoral fellow in the Department of Psychiatry and Biobehavioral Sciences at UCLA's David Geffen School of Medicine and the paper's first author.

Marco Iacoboni

A professor in the Department of Psychiatry and Biobehavioral Sciences at the David Geffen School of Medicine and a senior author on the paper.

UCLA Health

The academic medical center that conducted the study on the limitations of current AI systems.

Got photos? Submit your photos here. ›

What they’re saying

“While there is a current focus in world modeling on external embodiment, such as our outward interactions with the world, far less attention is given to internal dynamics, or what we term 'internal embodiment'. In humans, the body acts as our experiential regulator of the world, as a kind of built-in safety system.”

— Akila Kadambi, Postdoctoral fellow

“By contrast, current AI systems have no equivalent mechanism. They process inputs and generate outputs without any persistent internal state that regulates how they behave over time. This is not just a performance limitation, but also a safety limitation. Without internal costs or constraints, an AI system has no intrinsic reason to avoid overconfident errors, resist manipulation or behave consistently.”

— Marco Iacoboni, Professor

What’s next

The authors propose a 'dual-embodiment framework' to guide future research on building AI systems that model both their interactions with the external world and their own internal states, which could help address the safety and trustworthiness concerns raised in the study.

The takeaway

This study highlights a crucial gap in current AI technology - the lack of 'internal embodiment' or an awareness of one's own internal states, which humans rely on to regulate behavior. Addressing this limitation could be key to developing AI systems that are genuinely aligned with human experiences and behavior, rather than just superficially fluent.