The Godfather of AI wants an FDA-style approval system for the technology

• Geoffrey Hinton proposed an FDA-style approval system for AI models.
• Hinton said companies should prove their models are safe before releasing them, as drugmakers do.
• Hinton warns that recursive self-improvement could make AI risks escalate quickly.

Geoffrey Hinton wants AI companies to clear a hurdle familiar to drugmakers.

On Tuesday’s “Smart Girl Dumb Questions” podcast, the Nobel Prize-winning computer scientist, often called the “Godfather of AI,” proposed that AI companies convince a regulator that their products are safe before releasing them.

“You’re not allowed to just make a new drug and release it on the market,” Hinton told podcast host Nayeema Raza. “You have to convince the FDA. And to do that, you have to do a lot of work, about a billion dollars worth of work.”

“That seems like the very least we should have for AI,” Hinton added.

Hinton’s prescription for the AI safety debate came after predictions of AI catastrophes and after OpenAI’s decision to shelve its latest model, following testing that raised concerns about whether it would follow users’ instructions and stay within its authorization.

OpenAI’s head of safety systems, Saachi Jain, previously told Business Insider that GPT-6.1 Astra improved at persisting through difficult tasks but fell short on permissions and communicating what it had done.

Hinton said that AI is increasingly helping itself improve, potentially accelerating both its capabilities and its dangers.

“Exponentials grow very fast,” Hinton said. “We’re beginning to get recursive self-improvement because AI is being used to make AI better in many different ways.”

“So, it’s just a guess,” Hinton added, “But I guess we have a year or two before everything gets much worse than it is now.”