Anthropic researcher quits over AI fears, warns 'it could kill us all'

Former Anthropic researcher Jacob Coxon said in a resignation post on X that he and others at the company “earnestly believe” AI could kill humans by the end of the decade. The 27-year-old shared concerns that future systems could become superhuman, gain real power and resources, and hack widely. His comments gained traction after Anthropic employee Evan Hubinger agreed publicly and said he personally estimates the risk of superintelligence within 10 years at under 10%. Hubinger later pointed to Anthropic’s Risk Report, which says current AI poses little chance of achieving that level of control. Anthropic said it directed CNN to Hubinger’s follow-up.