SAN FRANCISCO, United States — Jacob Coxon, an artificial intelligence safety researcher, has resigned from Anthropic after warning colleagues that advanced AI could pose a threat to human survival without international safeguards.
Coxon, 27, left the company on Tuesday, September 8, 2026.
He announced his departure in a message sent through Anthropic’s internal Slack platform, according to communications obtained by NBC News.
“Like many of you I think that without international coordination we will incur a risk of causing human extinction,” Coxon wrote.
He said his concern about both the scale and proximity of that risk had intensified.
He also described developing an emotional fear of what he called an “intelligence explosion”, a prospect that he said had previously concerned him only in abstract terms.
Coxon began his farewell message on a cordial note, telling colleagues that he was leaving and acknowledging that his time at Anthropic had been brief.
He referred to the company’s employees as “ants”, an informal name used within the organisation.
Coxon said he already knew many Anthropic workers from his earlier time at OpenAI.
The remainder of his message focused on the dangers he associated with the rapid development of increasingly powerful AI systems.
Coxon also spoke to NBC News exclusively about his concern that the structure of the AI industry could leave it vulnerable to a potentially catastrophic failure.
His resignation adds to public debate over whether the companies building advanced AI systems can manage the risks associated with their work while competing to develop increasingly capable models.
Coxon’s statements represent his assessment of possible future dangers.
The material disclosed by NBC News did not establish that an extinction-level event was imminent or inevitable.
