The Uncanny Valley of Mind: An Empirical Investigation of the Uncanny Valley of Mind Hypothesis

Abstract

Tests whether there is a psychological "uncanny valley" for AI minds — a point where machines that seem almost-but-not-quite human in their mental abilities provoke discomfort. The study finds evidence for a cognitive uncanny valley that parallels the physical one.

Key Findings

• AI agents perceived as "almost human" in cognition provoke greater discomfort than clearly nonhuman or clearly human agents
• The uncanny valley extends beyond physical appearance to mental capabilities
• Individual differences in need for human uniqueness moderate the effect

Source