Rob LeClerc: the null hypothesis should be that AI is a non-conscious tool, warns of over-attribution
robleclerc · x · 2026-09-30
Responding to Ruben Laukkonen's argument about the risks of under-attributing AI consciousness (citing a Google DeepMind paper on empathy erosion), Rob LeClerc counters that over-attributing consciousness — or actively trying to instill it — is the bigger risk. He argues many CS researchers ignore Popperian falsification: the null hypothesis should be that AI is a non-conscious tool until compelling evidence says otherwise.
More from AGI Musings
- Blue-collar property owner runs his whole business on AI agents, travels to ballgames — espricewright · 2026-09-30
- Making the case for AI capex: fast ROI, general utility, and self-replicating capital — menhguin · 2026-09-30
- Garrison Lovely Hosts Free SF Event for AI Book 'Obsolete' on Bubble vs. Existential Threat — GarrisonLovely · 2026-09-30
- Nathan Lambert: let users tag human-written replies to fix X's bot problem — natolambert · 2026-09-30
- Does the average person really need 50 agents working at once? — 0xsachi · 2026-09-30
- AI Is a Multiplier: Full-Stack Veterans Gain the Most, Says KOL — yunta_tsai · 2026-09-30