A researcher at Anthropic has resigned and issued a stark warning about the future of artificial intelligence. Jacob Coxon, who previously worked at OpenAI before joining Anthropic, said he left the company because he believes AI labs are dangerously racing to build self-improving systems that they cannot properly control or understand.
In a public statement, Coxon said that he and others earnestly believe advanced AI could pose an existential threat to humanity. He accused major AI companies of moving too quickly toward superhuman systems without putting adequate safety measures in place first. He called for pacing agreements between competing labs to slow down development before the technology becomes too powerful to manage or contain.
The resignation is significant because Anthropic has carefully built its reputation around AI safety. Unlike some competitors, the company has publicly emphasized careful, responsible development and has attracted top researchers who specifically wanted to avoid reckless speed. When one of its own safety-minded researchers quits over these exact fears, it raises serious questions about whether even the most cautious labs are keeping pace with the risks they claim to take seriously.
The debate over AI safety has been growing louder across the tech industry and in government halls, but this departure adds real insider weight to the warnings. It comes from someone with direct experience training the very models that are becoming more capable each month. In the coming weeks, industry watchers expect this to fuel more calls for strict government oversight, independent audits, and binding safety rules. Whether rival labs will actually agree to slow down remains doubtful, but the pressure to prove that their powerful systems are truly safe is now higher than ever.