Will AI kill all humans? Anthropic employees freak out!

Former and current employees at Anthropic have publicly raised alarming concerns about advanced AI potentially causing catastrophic harm. Jacob Coxon, who resigned from the company, wrote on X that people building AI privately believe it could kill all humans by the end of the decade, while public messaging may sound less severe. Evan Hubinger, a current Anthropic staffer, replied that the company “earnestly” believes AI could kill all humans and that the alignment plan for superintelligence is not yet in place, estimating a “>10%” chance within ten years. The debate sparked comments from Florida Gov. Ron DeSantis, who urged sober guardrails rather than panic.