Former OpenAI safety lead calls for nuclear-style AI oversight

A former OpenAI safety lead who oversaw the lab's published safety reports compared frontier AI oversight unfavorably to nuclear power plants, warning of a "broken" internal culture.