DeepMind CEO Demis Hassabis Calls for Independent Standards Body to Regulate Frontier AI

DeepMind CEO Demis Hassabis proposed the creation of an independent standards body to oversee frontier AI model releases, drawing a direct parallel to the Financial Industry Regulatory Authority (FINRA). In a post on X on Tuesday morning, Hassabis outlined a framework that would require frontier labs to voluntarily submit models for review up to 30 days before release, with the goal of establishing mandatory safety testing for deployment in the U.S. market.

How the Proposed Standards Body Would Work

Under Hassabis’s plan, the new organization—backed by the U.S. government but funded by the AI industry—would test frontier models and develop best practices for their release. The system builds on ad hoc reviews previously conducted by the U.S. government on Anthropic’s Mythos and OpenAI’s Sol, which drew criticism for lacking technical expertise and transparency. Hassabis argues that a dedicated body could hand off those decisions to technical experts, ensuring consistent, informed oversight. The framework also includes provisions for addressing critical post-release vulnerabilities.

Political and Industry Context

The proposal arrives amid ongoing controversy over AI regulation within the tech industry and the Trump Administration. White House AI advisor Sriram Krishnan recently dismissed the possibility of an AI regulator within the executive branch, stating flatly that “there will not be an FDA for AI.” By modeling the new body as a self-regulatory organization like FINRA, Hassabis aims to address those concerns while still providing meaningful oversight. The standards body would be staffed by open-source representatives and technical experts from within the industry, with financial backing from AI labs to retain top talent.

Why This Matters for AI Development

Hassabis emphasizes that the approach is designed to keep pace with rapid AI advancement, adapting to emerging risks without stifling innovation. The regulator could outsource some evaluations to specialized AI safety groups, allowing for focused assessment of specific risks. “The strength of this approach is it would be technically focused, while at the same time supporting innovation and incentivising responsible behaviour,” Hassabis wrote. He added that the system could be “ratcheted up if the seriousness of the situation demands.” For readers, this represents a potential shift toward more structured, expert-led governance of the most powerful AI systems, balancing safety concerns with industry growth.

Conclusion

Hassabis’s call for a FINRA-style standards body marks a significant moment in the ongoing debate over AI regulation. By proposing a technically focused, industry-funded, but independent regulator, he aims to bridge the gap between calls for safety and resistance to heavy-handed government oversight. The success of such a framework will depend on political will, industry cooperation, and the ability to attract and retain technical expertise—but it offers a concrete path forward for governing frontier AI.

FAQs

Q1: What is the proposed AI standards body modeled after?
A1: It is modeled after the Financial Industry Regulatory Authority (FINRA), a self-regulatory organization that oversees financial firms in the U.S.

Q2: Who would fund and staff the new body?
A2: The body would be funded by AI labs but operate independently, with staffing from open-source representatives and technical experts within the industry.

Q3: Why is the proposal controversial?
A3: AI regulation remains divisive, with the Trump Administration opposing a formal regulator. Hassabis’s self-regulatory model aims to address those concerns while still providing oversight.