Paper Proposes 'Embodied Hijack' Hypothesis for Human Anthropomorphism of AI

MacrinePhD · x · 2026-08-11

A new paper introduces the Embodied Hijack hypothesis to explain why humans persistently attribute agency to conversational AI, even when they know the systems lack an inner life.

The authors argue that across evolutionary time, fluent communication and contingent responsiveness were exclusively produced by embodied, self-maintaining agents with vulnerability and temporal continuity. Current LLMs are the first entities to reproduce these signals without these grounding properties. This misalignment between evolved human cognition and disembodied AI leads to a predictable 'hijacking' of users' inferential systems, explaining the persistence of anthropomorphism.

Related event: Embodied Hijack Hypothesis Explains Anthropomorphism Towards AI(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →