I've been warning about this for some time now.

Last week, Anthropic AI researcher Jacob Coxon resigned, writing on X that OpenAI and Anthropic were "gambling with our lives," warning that "the people building AI earnestly believe that it could kill us all by the end of the decade." A lead Anthropic researcher, Evan Hubinger, agreed in a post on X that Coxon was right, saying, "we really do earnestly believe AI could kill all humans!"

They believe this could happen when AI gets too smart and takes control of things like power centers and nuclear weapons. This is basically the plot of the 1970 sci-fi film Colossus: The Forbin Project.