Google DeepMind CEO Demis Hassabis proposes US-led global AI standards body to review frontier models before release

Google DeepMind CEO Demis Hassabis has proposed a U.S.-led, industry-funded AI standards body modeled after FINRA to evaluate frontier models up to 30 days before release. Hassabis warns that artificial general intelligence is only a few years away, bringing imminent cyber, biological, and nuclear risks. The proposal follows recent ad hoc federal interventions that froze Anthropic's Mythos model and restricted OpenAI's GPT-5.6, highlighting the need for a structured regulatory framework.

US-led AI standards body

• ▪Google DeepMind CEO Demis Hassabis proposed a United States-led global AI standards body to evaluate frontier models before they are released
• ▪Demis Hassabis argued that the United States is best positioned to lead the AI standards body due to its economic and technical standing

FINRA-style regulatory model

• ▪The proposed standards body would be funded by the AI industry and operated independently under federal government oversight
• ▪The proposed AI standards body would be modeled after the Financial Industry Regulatory Authority, a private, self-regulatory organization that polices Wall Street

Thirty-day pre-release testing regime

• ▪Once the assessment protocol is proven effective, passing the safety review would become mandatory for deploying frontier models in the United States market
• ▪Under the proposal, frontier labs would initially voluntarily share models with the standards body for safety testing up to 30 days before release

AGI timeline warnings

• ▪Demis Hassabis warned that advanced AI models could soon present severe national security risks, including cybersecurity, biological, and nuclear threats
• ▪Demis Hassabis stated that artificial general intelligence is probably only a few short years away, creating a precious window to ensure safety

Mythos crackdown precedent

• ▪OpenAI agreed to restrict the initial rollout of its GPT-5.6 model to government-vetted partners following requests from the United States government
• ▪The United States government temporarily froze Anthropic's Mythos and Fable models under export controls, leading to 2.5 weeks of negotiations to release them

Industry leader support

• ▪Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have also called for United States-led frameworks or binding regulations to govern advanced AI models
• ▪Demis Hassabis has spent months quietly briefing the Trump administration, other AI labs, and European officials to build support for his regulatory proposal