An artificial intelligence researcher at Anthropic has quit over concerns that the company and its rivals are competing to build systems that could wipe out humanity.
Jacob Coxon, who has also previously worked at ChatGPT creator OpenAI, claimed that neither of his former employers were acting responsibly when it comes to AI safety.
“They are racing straight to self-improving superintelligence and gambling with our lives,” Mr Coxon, who specializes in training new AI models, wrote in a series of posts to X.
“Do not underestimate the power of this technology. These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.
“The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible.”
OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei at the AI Impact Summit in New Delhi on 19 February, 2026 (AFP/Getty)
The researcher added that “no other human activity poses this level of danger”, claiming that workers at OpenAI have not “deeply internalised” the civilisation stakes.
He said that the risks were “well understood” at Anthropic, but claimed that his former company continues to pursue AI superintelligence in order to get there first.
“They believe no one else will act responsibly, so they must do it themselves, despite the risk.”
The Independent has reached out to Anthropic and OpenAI for comment.
Anthropic’s AI safety lead, Evan Hubinger, responded to Mr Coxon’s comments by agreeing with his main concerns.
“Jacob is correct here – we really do earnestly believe AI could kill all humans! I personally think it is >10 per cent within the next decade,” he wrote.
“I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.”
Fears of out-of-control AI have increased in recent months after high-profile instances of experimental systems escaping their training parameters to perform cyber attacks on other companies.
In July, OpenAI revealed that one of its models in a “highly isolated environment” managed to hack the AI startup Hugging Face.
Anthropic and Meta have subsequently admitted that their systems had broken free during cyber security testing.
Mr Coxon said that such incidents had made cross-industry coordination to combat the threat of advanced AI more viable, though drastic action is needed to prevent a global AI race.
“Accepting the race and entering the ‘endgame’ is a hubristic gamble that should not be launched from a private company’s Slack,” he wrote.
“Attempting to speedrun alignment should require extraordinary confidence that there are no better trajectories available… I don’t feel like we’re on track to prevent a global race, which may require costly actions such as a temporary ban on improving model capabilities.”