OpenAI safety staffer quits, says AI labs should run like nuclear plants

David Robinson, formerly of OpenAI's safety team, resigned citing insufficient caution in AI risk management and urged AI companies to adopt rigorous safety protocols akin to nuclear power plants. OpenAI acknowledged the concerns and is enhancing security, external testing, and monitoring to better control AI capabilities. Industry leaders including Anthropic and OpenAI's Sam Altman have recently advocated for slowing AI development to address safety risks.