Anthropic’s Evan Hubinger: >10% chance AI kills everyone this decade, no alignment plan yet
Anthropic alignment researcher Evan Hubinger backs Jacob Coxon’s claim that AI builders earnestly fear extinction-level risk, putting his personal odds above 10% within a decade and saying Anthropic is trying but still lacks a plan to align superintelligence and is not clearly on track.
Spotted on X — source: Evan Hubinger