Anthropic researcher resigns over AI extinction fears

Jacob Coxon, a researcher at Anthropic, has resigned, citing fears that self-improving artificial intelligence systems pose a risk of human extinction. TechCrunch reported the resignation and Coxon’s stated reasoning on September 9.

According to the report, Coxon called for pacing agreements between AI labs as a response to the risks he identifies. The source record does not specify further detail on the structure or scope of such agreements, nor does it elaborate on the technical basis for his extinction concerns beyond the stated warning against self-improving AI.

The report’s headline characterizes the situation as “gambling with our lives,” though the supplied source material does not establish this phrase as a verified direct quotation from Coxon himself.

The departure is reported as occurring amid broader questions about the pace of AI development within the industry. The source material identifies Anthropic as the company from which Coxon resigned but does not detail the company’s response, if any, to his stated concerns or to his departure.

No further elaboration on Coxon’s prior role, tenure, or specific technical objections is available in the supplied source record. The report is attributed to TechCrunch AI and dated September 9, 2026.