Geoffrey Hinton proposes fda-style regulation for AI industry

On October 7, 2026, Geoffrey Hinton, known as the 'father of AI' and a Nobel laureate, suggested that AI companies should adopt a regulatory framework similar to the pharmaceutical industry's FDA approval process. In an episode of the podcast 'Smart Girl Dumb Questions,' Hinton emphasized that AI firms must prove their products are safe before release, akin to the rigorous testing required for new drugs. He estimated the cost of such validation could reach billions of dollars. Hinton's proposal comes amid growing concerns about AI's potential for catastrophic outcomes and recent pauses in model development by OpenAI due to safety concerns. Saachi Jain, OpenAI's safety systems lead, highlighted that while GPT-6.1 Astra has improved in handling complex tasks, it still lacks in permission controls and behavioral logging. Hinton warned that AI's self-improvement capabilities could accelerate its growth, potentially leading to a critical point within a couple of years.