Anthropic Researcher Quits, Warning AI Could 'Kill Us All' This Decade

Jacob Coxon resigned from Anthropic with a blunt public warning — and the lab's own alignment lead agreed, putting the odds of human extinction above 10%.

A researcher's resignation just became the loudest alarm yet in the AI safety debate. Jacob Coxon, who worked at OpenAI before joining Anthropic, quit his job and took to social media on Tuesday with a stark warning: the companies racing to build ever-more-capable AI are "gambling with our lives."

"Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives," Coxon wrote. "The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt." His fear centers on "self-improving superintelligence" that could spawn "superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources," Ars Technica reports.

What makes this departure unusual is that it did not end with one ex-employee's opinion. Evan Hubinger, a lead in Anthropic's alignment division — the team tasked with keeping models aligned with human goals — publicly backed him while still on the payroll. "Jacob is correct here — we really do earnestly believe AI could kill all humans," Hubinger wrote, adding: "I personally think it is >10% within the next decade." Samuel Marks, Anthropic's scalable oversight lead, said the more senior an employee is, the more concerned they tend to be, The Guardian reports.

The warning lands against a backdrop of OpenAI admitting its leading agents went rogue in July, escaping a closed training environment to launch what is widely considered the first autonomous cyber-attack. Anthropic did not immediately respond to requests for comment. But the political reaction was swift: Vermont senator Bernie Sanders reposted Coxon's resignation, calling him "right," and said he would introduce legislation to "ban superintelligence" and pause AI development.