AI: Incidents and Anthropic researcher's resignation push 'AIpocalypse' debate to new heights

It was an unusual scene. As BBC journalist Victoria Derbyshire interviewed artificial intelligence (AI) pioneer Geoffrey Hinton live on air on Thursday, September 10, she was left speechless, stunned by his words. When she asked, "Do you believe there is a greater than 10% chance that AI could kill all humans potentially within a decade," the 2024 Nobel Prize in Physics laureate replied, "A 10% chance does seem not an unreasonable estimate to me." The journalist froze. Silence. "Wow." Silence. "Oh my God."

This is the tone of the debate that has gripped the world of AI, and far beyond, since Wednesday, when Jacob Coxon, a 27-year-old Briton, announced publicly that, for ethical reasons, he was leaving Anthropic, the AI company where he worked for a few months after a stint at OpenAI. "Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives," he wrote. "The people building AI sincerely believe it could kill us all by the end of the decade. This is not a marketing stunt."

His message triggered an avalanche of reactions. It was viewed more than 160 million times on the social network X and generated over 18,000 comments, notably from peers, who supported the claims. "Jacob is correct here: We really do earnestly believe AI could kill all humans," said Anthropic safety lead Evan Hubinger, estimating the risk of that happening at "[more than] 10% within the next decade." Another employee at the company, Samuel Marks, agreed, specifying it "could happen in the next few years" and that "the more senior the employee, the more concerned they are."

You have 76.85% of this article left to read. The rest is for subscribers only.