Anthropic researcher says more than 10% chance AI could kill all humans after colleague quits

I urge you to go to X and read the threads of the conversation unfolding there after an Anthropic researcher announced his resignation last night, exposing that everyone in the company thinks that AI will “kill us all by the end of the decade,” and the Alignment Science lead at the company agreed. 🧵

An Anthropic researcher resigned, citing concerns about the destructive potential of AI systems and the conduct of the top companies working on the technology. "The people building AI earnestly believe that it could kill us all by the end of the decade," he said. www.huffpost.com/entry/live-u...

"more than 10% chance AI 'could kill all humans' " But they are going to carry on anyway Just like Oppenheimer, who thought the first atomic explosion could have a chain reaction and destroy the planet......but still went ahead Hubris and stupidity www.bbc.co.uk/news/article...

Explain to me again why governments are allowing AI to remain unregulated and left to the whim of billionaires who clearly do not care for our safety and futures Anthropic researcher quits saying AI developers believe 'it could kill us all' by the end of the decade. www.abc.net.au/news/2026-09...

I interviewed for a low six-figure comms job focused on AI safety once and they asked me what challenges I saw for explaining singularity / extinction risks and I said they needed examples of how advanced AI systems — on their own — could get to nukes or bioweapons. They were like “huh, okay.”

Me: "AI is bad because it steals art without permission or compensation, destroys the environment, and wrecks cognitive skills." Anthropic researcher: "Hold my beer..." www.bbc.co.uk/news/article...

The call is coming from inside the house. Safety researchers are resigning, powerful AI models are breaking out of their labs, and companies are racing ahead anyway. It's past time for Congress to get off the sidelines and do its job. We can start with my bipartisan FRONTIER Act.

There’s a certain poetic justice in all this: some of the world’s smartest psychopaths spending trillions to build systems that have a reasonable shot at eliminating ALL human life in the near future. www.cnbc.com/2026/09/09/a...

I bet the chance of a demented orange polyp in charge of the U.S. starting a nuclear war and killing us all is much higher. www.cnbc.com/2026/09/09/a...

Oh great BBC News - Anthropic researcher believes more than 10% chance AI 'could kill all humans' - BBC News www.bbc.co.uk/news/article...

I’m not even sure what we are supposed to do of these doomsday prophecies by the people who are bringing about doomsday machines Should we trust them? Deride them? Both? www.cnbc.com/2026/09/09/a...

It's ok to discuss that #AI might kill all humans. It is NOT ok to discuss that overpopulation and overshoot - aka eating animals, cars, airplanes, pets and making babies - will kill all humans. Why? Because humans are alienated (decoupled) from reality. #science of #extinction

Jacob Coxon, a former OpenAI and Anthropic researcher, resigned from Anthropic yesterday, warning: “The people building AI earnestly believe that it could kill us all by the end of the decade.”

AI researcher Jacob Coxon quitting Anthropic and warning that tech bosses privately believe self-improving superintelligence "could kill us all by 2030" is alarming. Prioritising commercial market dominance over human safety is reckless. When will we regulate Big Tech? (1/2) 🤔 #FBPE

Recent AI cyber attack events have shown that capabilities are advancing faster than our abilities to control them. Fierce competition is driving firms to release models fast. Even if they are safe under testing conditions, they aren't necessarily safe when deployed. www.bbc.co.uk/news/article...