Hassabis calls for US-led body to police frontier AI

TL;DR:

• Google DeepMind chief Demis Hassabis has called for the US to spearhead a standards body overseeing new AI models and assessing national security risks.
• He proposes a federally overseen public-private partnership modelled on FINRA, funded largely by industry.
• Labs would initially share models voluntarily up to 30 days before release, becoming mandatory for US deployment once the process proves effective.

Demis Hassabis wants the testing of frontier AI models placed under an American standards body. The Google DeepMind chief and Nobel laureate said in an article posted on X that “urgent action” is needed on risks tied to artificial general intelligence — the point at which AI matches or surpasses human intelligence — which he suggested is “probably only a few short years away”.

A regulator modelled on FINRA

The proposal is specific. Hassabis argued the US is well positioned to lead “given its economic and technical standing”, and that it “could establish a new Standards Body modelled on a federally overseen public-private partnership or self-regulatory organisation, much like the Financial Industry Regulatory Authority (FINRA), with a board that includes independent leading technical experts and open-source representatives”.

Funding would be “substantial” and would “likely” come from industry, he said — enough to attract technical talent and provide compute for large-scale testing. Frontier labs would initially share models voluntarily for review up to 30 days before release, becoming mandatory for US-market deployment once shown to be effective. Tests for agentic systems could look for attempts to bypass safety guardrails or signs of deception, alongside watermarking generated images.

“We’ve already seen the challenges frontier models pose for cybersecurity, and other threats including nuclear and bio risks may soon emerge as capabilities continue to advance,” he said.

For a UK audience the awkwardness is hard to miss. The head of Britain’s flagship AI lab is arguing that standards leadership belongs in Washington — on the same day the chancellor staked her growth argument on UK AI sovereignty and the Bank of England governor argued that oversight must be coordinated internationally rather than led by any single country. Britain hosted the first AI Safety Summit; its most prominent AI figure is now proposing the referee sit elsewhere.

Hassabis is not alone. Sources told CNBC he and Anthropic chief Dario Amodei called for a US-led coalition on AI rules at a G7 meeting last month, and OpenAI’s Sam Altman proposed a similar body in the Financial Times.

Looking forward

The proposal arrives amid friction rather than consensus: Anthropic was recently locked in negotiations after the Trump administration imposed temporary export controls on an advanced model, and OpenAI was asked to limit a rollout. Meanwhile Chinese releases from DeepSeek and Z.ai are gaining traction with US companies as AI costs rise, and US lawmakers are weighing how to curb their adoption. A standards body funded by the industry it polices will invite the obvious question about who it actually serves.