AI developers warn their own technology could escape human control

Artificial intelligence developers are warning that their own technology could escape human control, with a former OpenAI and Anthropic researcher stating colleagues genuinely believe AI may kill everyone by the decade's end. Jacob Coxon quit the industry on September 9, 2026, accusing both companies of racing toward self-improving superintelligence and gambling with human lives. His post, viewed over 150 million times, was publicly corroborated by Anthropic safety lead Evan Hubinger, who estimated over ten percent probability of AI destroying humanity within ten years. Coxon's departure differs from typical quiet exits, as his warning gained unprecedented traction. The comparison to boarding a plane with a captain announcing a crash probability above ten percent highlights the industry's ongoing race to build increasingly capable models despite these acknowledged existential risks.