Worried Anthropic researchers warn that AI ‘could kill all humans’ | The Verge
Worried Anthropic researchers warn that AI ‘could kill all humans’ | The Verge
摘要: AI
Worried Anthropic researchers warn that AI ‘could kill all humans’
A departing researcher says AI companies are ‘gambling with our li
正文
Worried Anthropic researchers warn that AI ‘could kill all humans’
A departing researcher says AI companies are ‘gambling with our lives’ by racing to develop self-improving AI.
A departing researcher says AI companies are ‘gambling with our lives’ by racing to develop self-improving AI.
Robert Hart
Sep 9, 2026, 9:56 AM UTC
Image: The Verge
Robert Hart
is a London-based reporter at
The Verge
covering all things AI and a Senior Tarbell Fellow. Previously, he wrote about health, science and tech for
.
A senior Anthropic safety researcher has said there is more than a 10 percent chance artificial intelligence “could kill all humans” by the end of the decade, just hours after a colleague resigned over fears the AI lab and its rivals are carelessly racing to build “superhuman systems” they cannot control.
on X announcing his departure, Jacob Coxon, a researcher who has trained AI systems at Anthropic, said he had quit the company over its lax approach to safety. Coxon, who previously trained systems for OpenAI, accused the two AI companies of “racing straight to self-improving superintelligence and gambling with our lives,” even though “the people building AI earnestly believe that it could kill us all by the end of the decade.”
Industry insiders have long expressed concerns about the potential dangers of self-improving AI systems, which they warn could spiral out of human control in a runaway loop often described as recursive self-improvement. While not yet realized, companies are actively pursuing this goal and much of today’s AI code is written with the help of AI.
In a direct response, Evan Hubinger, who leads one of Anthropic’s AI safety teams,
he worries about self-improving AI, adding that it “is happening faster than we thought.” He also agreed with Coxon’s characterization. “We really do earnestly believe AI could kill all humans,” he said, personally estimating the chances to be gre