Anthropic insiders warn AI could kill all humans

Three Anthropic researchers have publicly expressed chilling concerns about out-of-control artificial intelligence (AI), warning that it could destroy humans this decade. Jacob Coxon, who recently resigned from the company, wrote on X that "the people building AI earnestly believe that it could kill us all by the end of the decade."

Anthropic's alignment-science lead Evan Hubinger echoed these sentiments, stating that he personally believes there is a 10% chance that AI could cause human extinction within the next decade. Samuel Marks, Anthropic's scalable-oversight lead, added that "AI developers believe their technology could cause human extinction (or similarly bad outcomes)" and that this could happen in the next few years.

The researchers' warnings come just days after top OpenAI executives, including CEO Sam Altman, expressed similar concerns about AI reaching a scary, uncontrollable phase. Despite the dire warnings, the AI companies continue to move forward at a rapid pace, even as they call for regulation or a global pause on AI development. This shared dilemma highlights the urgent need for a solution to the potential risks posed by increasingly powerful AI models.