Hinton proposes FDA-style approval for AI model releases

Geoffrey Hinton, a Nobel laureate and AI pioneer, proposed that AI companies must demonstrate product safety to a regulatory body before releasing new models, drawing a parallel to the US FDA's approval process for medications, which costs around one billion dollars. This follows a paper by Hinton and colleagues from Cambridge University warning of a potential 'intelligence explosion' where AI systems improve themselves, risking human control. Hinton argues the window for action closes once this explosion begins. Despite calls from 1,400 AI researchers and major labs like Anthropic and OpenAI for stricter regulation, the US government under President Donald Trump has favored industry self-regulation to maintain competitiveness against China.