• 2 min read
DeepMind chief pushes new AI standards body
DeepMind CEO Demis Hassabis wants an independent, FINRA-style body to review frontier AI models before release in the U.S.
Image: TechCrunch
Google DeepMind CEO Demis Hassabis is calling for a new independent body to oversee releases of frontier AI models in the U.S. In a Tuesday morning post on X, titled “A Framework for Frontier AI and the Dawning of a New Age,” Hassabis argued for a standards body modeled on the Financial Industry Regulatory Authority (FINRA).
Under the proposal, frontier labs would voluntarily share models up to 30 days before release so the body could test them and develop best practices. If that review process proved effective, Hassabis said it could later be formalized, requiring frontier models to pass the assessment before being deployed in the U.S. market. Labs would also work with the group on any critical vulnerabilities found after release.
The idea builds on the ad hoc U.S. government reviews of Anthropic’s Mythos and OpenAI’s Sol. Those reviews faced criticism over limited technical expertise and opaque decisions about when a model could launch.
Hassabis suggests handing those judgments to a new organization that would be backed by the U.S. government, funded by the AI industry, and operated independently. That structure could help navigate resistance to direct AI regulation from both the tech industry and the Trump administration. Sriram Krishnan, a White House AI advisor and a16z general partner, recently rejected the idea of an executive-branch AI regulator, saying there “will not be an FDA for AI.”
Recommended reading
Open models are overtaking AI’s front line
Hassabis said the proposed body could include open source representatives and technical experts from the industry, with enough financial support from AI labs to retain them. It could also outsource parts of its evaluation work to specialist AI safety groups.
“The strength of this approach is it would be technically focused, while at the same time supporting innovation and incentivising responsible behaviour. It is designed to keep up with the field’s acceleration and adapt to the biggest risks as they are identified, and could be ratcheted up if the seriousness of the situation demands.”
AI Editor
Ava covers the rapidly evolving world of artificial intelligence, from foundational models and research labs to the real-world economics of intelligence. With a background in computational linguistics, she cuts through the hype to find out what actually works. She firmly believes that benchmarks are just marketing until reproduced in the wild.
via TechCrunch
