Researcher Pushes Back on Harari's Trust-Shifts-to-AI Thesis With Four Empirical Studies

AryHHAry · x · 2026-09-25

Responding to Harari's claim that trust is shifting from humans to AI, a researcher cites four studies showing the picture is messier: trust calibration (not one-way shift) is what matters — overtrust causes misuse, undertrust disuse (Lee & See, 2004); algorithm aversion means people abandon algorithms faster than humans even when the algorithm is more accurate (Dietvorst et al., 2015); empathy labeled as AI is rated lower (Nature Human Behaviour, 2025); and people trust and cooperate less with LLM opponents (PNAS Nexus, 2025).

Original post →

More from AGI Musings

AGI Musings channel →