AI could destroy human race over the next decade, three Anthropic researchers warn

Three Anthropic researchers are claiming that out-of-control AI could destroy humans this decade, and the trio aired their concerns on X. 

Jacob Coxen was an AI researcher with Anthropic before he resigned from the company on Tuesday, citing unacceptable risks that weren't being considered in the race to develop AI. 

In a series of posts on X, Coxon claimed that there will soon be "superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources." 

"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 warned.

He said while the fears are not aired publicly, the same people promoting AI express these concerns privately. He said the development of AI continues because at OpenAI, one of Anthropic's competitors, "many have not internalized the civilization stakes." 

Anthropic alignment-science lead Evan Hubinger backed up Coxen's claims and added, "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."

Samuel Marks, Anthropic scalable-oversight lead, chimed in with: "AI developers believe their technology could cause human extinction (or similarly bad outcomes). This could happen in the next few years. In general, the more senior the employee, the more concerned they are." 

Marks said AI development continues despite the risks as a result of a "mixture of commercial incentives and a belief that they are in a race with each other." 

None of the researchers provided specifics on how this extinction event would occur.