Reading AI outputs is talking to yourself: anthropomorphism is a cognitive bias

gerardsans · x · 2026-11-20

gerardsans argues that anthropomorphism — projecting human traits onto AI — is a documented cognitive bias about the observer, not the machine. AI is a closed symbolic system with no access to reality or experience, so reading AI responses means talking to yourself. AI labs deliberately added an interlocutor that works like a rubber duck — a design choice, not evidence of a self; companies have no incentive to disclose the anthropomorphism triggers, which are now banned in contexts like AI companions.

Related event: Developer Argues AI Anthropomorphizing Is a Reader's Cognitive Bias(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →