Anthropic staff warn AI could kill us all; RAND calls it faith

A former Anthropic researcher resigned, posted that the people building AI "earnestly believe that it could kill us all by the end of the decade," and the post picked up more than 100 million views in 24 hours. Nature's news explainer by Elizabeth Gibney, published 22 September 2026, walks through what those fears are, who is voicing them, and what evidence sits behind them.

The people making the loudest claims are not outsiders. Jacob Coxon told the Wall Street Journal on 8 September why he had quit. Evan Hubinger, Anthropic's alignment lead, puts the risk of human extinction at ">10% within the next decade." Dario Amodei, the company's chief executive, has called for a slowdown, but not a halt, in AI development. Around 1,400 employees across the AI sector have signed an open letter asking for the same.

Researchers outside the labs are less convinced. Michael Vermeer of the RAND Corporation says the extinction debate "involves so many untestable claims that you just end up with a conversation that is really more like faith than something scientific or empirical." Heidy Khlaaf of the AI Now Institute calls the framing "fear-mongering" and would rather regulators focus on AI's "low reliability and accuracy rates in critical environments with life-or-death consequences."

Four analysts we track had already surfaced the piece by the time it hit our inbox.

Nature does not pick a winner. It also notes that the labs calling for a slowdown stand to gain from it: stricter regulation can protect incumbents from competition and from liability for model-enabled attacks.