Jacob Coxon, a former lead safety researcher at Anthropic, told the BBC that employees at the company were "seriously alarmed" by the pace of artificial intelligence development and the risks it poses to humanity. Coxon previously worked at OpenAI and was responsible at Anthropic for training AI systems.
He left Anthropic a few days before giving the BBC interview. In a post on X (formerly Twitter) explaining his departure, Coxon wrote that he believes the company does not take safety seriously. In his BBC remarks he also said he thinks OpenAI and Anthropic are "competing to develop self-improving superintelligence," and that this competition is "playing with our lives."
"I think if we don't slow the current rate of progress, there is a good chance we will all die in the near future," Coxon told the BBC. He is 27 years old.
Coxon's departure follows Dario Amodei, Anthropic's CEO, publicly calling for a slowdown in AI development. Amodei proposed extending a pause-like regulatory approach across the industry and instituting independent oversight of AI model development. Sam Altman, CEO of OpenAI, and Elon Musk have both expressed agreement with that call.
Coxon said he supports such regulation but argued to the BBC that any measures must also be enforced with Chinese developers to avoid an international race. He added that people working at these companies are serious about calling for regulation because they feel forced into a competition and fear its consequences.
Coxon also discussed the hardest question: what an AI-caused apocalypse would look like. One risk Amodei has highlighted is a swarm of bots acting like a supercomputer that takes control of the internet. Coxon believes that scenario could become reality within as little as six months, and at the latest within a year.
Why this matters
The account matters because a senior safety researcher in the field publicly expressed alarm, left his employer over safety concerns, and gave a concrete — though alarming — timeline for severe outcomes. It also underlines that several prominent figures in the industry, including Dario Amodei, Sam Altman and Elon Musk, have supported calls for slowing development and applying independent oversight to advanced AI.
Coxon's statements raise questions about the need for international coordination and how to create effective, consistent regulation across different countries and companies.
Key facts
- Jacob Coxon is 27 years old, previously worked at OpenAI, and served at Anthropic as a lead safety researcher responsible for training AI systems.
- He left Anthropic a few days before speaking to the BBC and said on X that he left because the company did not take safety seriously.
- Dario Amodei has proposed slowing AI development and independent oversight; Sam Altman and Elon Musk have expressed agreement with that proposal.
- Coxon warned that an extreme scenario, such as a bot swarm taking over the internet, could occur within six months to a year.
