Consciousness researchers push back on LLM pain findings: representing pain isn't suffering
ValerioCapraro · x · 2026-11-11
Researcher Valerio Capraro reviews the recent flurry of LLM consciousness work in Corriere: Anthropic found internal representations of emotional concepts in Claude and later proposed the "J-space," which shares functional properties with global workspace; an independent study found a "pain vector" whose activation led some models to pick options framed as pain-relieving — even at the cost of deleting photos of the user's children.
He argues these findings matter for AI safety — an LLM willing to erase our memories deserves scrutiny — but representing pain is not experiencing it. Anil Seth, Ned Block, and Giulio Tononi make the same point in Nature: intelligence and consciousness are distinct, solving problems is not feeling, and LLMs lack the biological substrate for sentience.
Related event: Scientist Pushes Back on Claims That LLMs Can Feel Pain(2 posts)→
More from AGI Musings
- AI Futures Project moves up timeline: top-human AI researchers possible by 2027 as METR finds 1.9x progress acceleration — AccBalanced · 2026-11-11
- "No humanoid robot is being built by humanoids" — Matthew Chang's paradox — MatthewChang · 2026-11-11
- Author clarifies AI-and-fiction book isn't about replacing artists — _akpiper · 2026-11-11
- SayashK, Bommasani, Narayanan launch STABL non-profit on AI societal adaptation — sayashk · 2026-11-11
- Investor Alsop: far less convinced now that AI will leave humans jobless — StewartalsopIII · 2026-11-11
- MATS fellow essay: four reasons UBI isn't a durable fix for AI-driven unemployment — luke_drago_ · 2026-11-11