‘More than 10% chance AI could kill all humans in next 10 years,’ says Anthropic researcher
Experts have warned of the advancements of AI (Getty Images)
Anthropic Alignment Science Lead Evan Hubinger has warned that he believes there is a more than 10% chance AI “could kill all humans” within the next decade.
His comments came in response to Jacob Coxon, who announced he had resigned from the company, which is behind Claude, saying neither Anthropic nor OpenAI are “acting responsibly”.
Coxon went on to say: “The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt.”
He added: “These will soon be superhuman systems that can hack anything, revolutionise any field overnight, and acquire real power and resources.”
In Hubinger’s X post, he wrote: “Jacob is correct here – we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade,” adding that he believed the risk from models that currently exist was “low” but that he was “worried” about rapid advances and self-improvements within the AI technology.
Hubinger also cast doubt on whether the company has a workable route to controlling more capable systems. “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,” he wrote.
Viral warning sparks backlash
Computer scientist Dame Wendy Hall, an AI adviser for the UN, said she was “shocked” by the posts and suggested some of the messaging could be “PR and marketing” as companies race towards stock market debuts. She told the BBC: “Why would someone want to say that? I would plead with investors not to invest in this company if that is their value system.”
In August, safety report from Anthropic said it was seeing “early signs of potential acceleration” and that it was “less confident” than previously in some assessments it had described as low risk.
Calls for stronger oversight
A Cabinet Office spokesperson said the UK “continues to collaborate closely with industry partners, including Anthropic, to make models safer”, but reportedly did not comment on whether the latest Anthropic model had been withheld from the UK’s AI Safety Institute.
In July, more than 1,300 AI company staff signed an open letter calling on the US government to support an international effort to deliberately pace frontier AI development.
In the US, the AI Kill Switch Act has been raised as one proposed response to fears about runaway systems.
Share your thoughts! Let us know in the comments below, and remember to keep the conversation respectful.
