Anthropic CEO Dario Amodei published an essay on Saturday calling for the AI industry to slow the pace of frontier development, days after pretraining researcher Jacob Coxon's public resignation triggered the most sustained internal-safety reckoning at a major AI lab this year6,7,8.
Sam Altman of OpenAI and Elon Musk of xAI both publicly endorsed Amodei's proposal for industry-wide deceleration, regulation, and independent monitoring of AI model development. OpenAI chief scientist Jakub Pachocki separately called for "extreme caution" on Sunday3.
Three researchers, one message
The chain of events began on Tuesday, September 8, when Coxon, a 27-year-old British researcher who spent three years on pretraining work at OpenAI and then Anthropic, resigned and posted on X that both companies are "racing straight to self-improving superintelligence and gambling with our lives"2,4. He left Anthropic after four months, two months before his equity would have vested9.
Within hours, two colleagues still employed at Anthropic went public with similar warnings. Evan Hubinger, Anthropic's Alignment Science lead, wrote: "Jacob is correct here — we really do earnestly believe AI could kill all humans. I personally think it is 10% within the next decade". Hubinger added that Anthropic "does not yet have a plan to solve alignment for superintelligence and is not clearly on track to". Samuel Marks, who leads scalable oversight at Anthropic, said the people closest to the technology tend to become more concerned as they gain seniority.
Coxon told the BBC's Laura Kuenssberg on Saturday that people working on AI are "genuinely frightened" and that "if we don't slow down at the current rate of progress, there is a strong chance that we could all die in the immediate future"11,12. He said colleagues were contemplating buying rural land out of fear of instability from rapid AI progress. He estimated a swarm-of-bots scenario capable of hijacking parts of the internet could become realistic within six months to a year.
Industry fault lines
Not everyone agreed. Nvidia's Jensen Huang reportedly dismissed Coxon's warnings as untrue at a Goldman Sachs conference, consistent with his prior characterization of AI-ending-humanity claims as "complete nonsense". Hugging Face CEO Clement Delangue compared asking Coxon about extinction risk to asking an air conditioning technician about climate change. Some industry figures suggested the warnings could be designed to build hype ahead of potential IPOs by Anthropic and OpenAI.
Anthropic said it "has always been transparent that AI will bring both enormous benefits and unprecedented risks" and that it "continues to build models with some of the strongest safeguards in the industry"5. The company said the industry would benefit from "a lawful, verifiable way to work together to pace how powerful models are released".
Regulatory momentum
The episode has already intersected with legislative action. Senator Bernie Sanders and Representative Greg Casar introduced the Ban Artificial Superintelligence Act on September 3. British Labour MP Alex Sobel introduced the Artificial Superintelligence Security Bill in Parliament on Tuesday. California Governor Gavin Newsom signed two AI safety-evaluation bills backed by both OpenAI and Anthropic shortly after Coxon's posts[2].
Coxon said any slowdown would need to be coordinated with China to prevent an international race. He noted that workers at AI companies "are completely serious when they ask for regulation because they find themselves trapped in a race. And they're scared of the outcomes of that race".
ANALYSIS The convergence of Amodei, Altman, Musk, and Pachocki around a slowdown message within days of Coxon's resignation marks the first time the heads of three competing frontier labs have simultaneously endorsed deceleration. The open admission by Anthropic's own alignment lead that the company lacks a superintelligence-alignment plan puts concrete technical substance behind what had been a largely rhetorical debate.
Geoffrey Hinton told the BBC that a 10% chance of AI killing all humans "was not unreasonable".