DeepMind CEO proposes independent body to regulate frontier AI

On Tuesday morning, Google DeepMind CEO Demis Hassabis proposed creating an independent standards body modeled after the Financial Industry Regulatory Authority (FINRA), the US financial self-regulatory organization. The body could test frontier models and develop best practices for their release. Under the initial phase of the framework, Frontier Labs would voluntarily share models with the Standards Body for review up to 30 days before release. Once the assessment protocol is shown to be effective, formalization could follow quickly, meaning frontier models would be required to pass it to be deployed in the US market. Labs would also work with the Standards Body to address any critical post-release vulnerabilities.

The proposed system would build on the ad hoc reviews the US government performed on Anthropic’s Mythos and OpenAI’s Sol. Those reviews drew criticism for lack of technical expertise and opaque decision-making as to when a model could be released. Under Hassabis’s proposal, those calls could be handed to a new organization — backed by the US government but funded by the AI industry and operated independently, staffed by open-source representatives and technical experts drawn from within the industry.

The push for self-regulation comes as prospects for a dedicated government AI regulator dim. White House AI advisor and Andreessen Horowitz (a16z) general partner Sriram Krishnan recently said, “there will not be an FDA for AI.” Structuring the standards body as a self-regulatory organization, similar to FINRA, could be one way to address the resulting gap. Hassabis envisions the body funded by AI labs, which would provide the resources needed to retain technical staff, and argues it could even outsource some evaluations to the growing pool of specialized AI safety groups equipped to focus on particular risks.

Why it matters

As executive-branch AI regulation faces political headwinds in the US, industry leaders are attempting to preempt heavy-handed government oversight by proposing self-regulatory frameworks that lean on private technical expertise rather than a new federal agency.