An Anthropic researcher named Jacob Coxon resigned on September 9, 2026, warning that artificial intelligence could kill all humans by the end of the decade.
He said people building advanced AI models earnestly believe the technology could cause mass human extinction within ten years.
Coxon noted that this warning is not a marketing stunt or exaggeration.
The senior staff member estimated the probability of AI causing human extinction within a decade at more than 10%.
Other staff members noted that senior developers tend to hold deeper concerns about unconstrained recursive self-improvement and safety.
Evan Hubinger says Jacob is right, but the risk is low with the current model.
He said it’s low, but not nothing, and we are talking about destroying all humans.
Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. 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. https://t.co/QAIHiFP3QZ
— Evan Hubinger (@EvanHub) September 9, 2026
Hubinger continued. “What I am worried about is superintelligence arising from recursive self-improvement, as we have said is happening faster than we thought.”
Shouldn’t he declare war on those in charge?
The Godfather of AI agrees it is more than 1% dangerous.
“You just said that 10% doesn’t seem an unreasonable estimate that AI could kill all humans”
“Yes”
“Wow… oh my God.”@vicderbyshire speaks to Nobel Prize Winner and so-called ‘Godfather of AI’ Geoffrey Hinton about predictions by an Anthropic researcher that there’s a more… pic.twitter.com/sGYQI57REF
— BBC Newsnight (@BBCNewsnight) September 9, 2026
I don’t know if it will kill us all in ten years or ever, but we can’t stop it. We can’t let our enemies develop it and leave us nothing to fight them with.
We should have strict rules about how the scientists operate, or we could end up with something far worse than pandemics.
Maybe a rogue AI will get me in the end, but fortunately I am very unimportant and will be able to hide out. And I will always be very nice to my computer. I am well on my way and am very polite to Siri and Alexa.