On Tuesday morning, Google DeepMind CEO Demis Hassabis took to X to advocate for the establishment of a new regulatory body dedicated to overseeing frontier model releases. In a post titled "A Framework for Frontier AI and the Dawning of a New Age," he proposes a "standards body" modeled after the Financial Industry Regulatory Authority (FINRA). This entity would be responsible for testing frontier models and developing best practices for their deployment.

The post outlines a process where "Frontier Labs would voluntarily share models with the Standards Body for review up to 30 days before release." Once the assessment protocol proves effective and robust, formalization could follow swiftly, meaning frontier models would need to pass this review to be deployed in the US market. Labs would also collaborate with the Standards Body to address any critical post-release vulnerabilities.

This proposed system builds on the ad hoc reviews conducted by the US government on Anthropic's Mythos and OpenAI's Sol. Those reviews faced significant criticism for lacking technical expertise and for opaque decision-making regarding when a model could be released. Under Hassabis's vision, these decisions would be delegated to a new organization, backed by the US government but funded by the AI industry and operated independently.

The prospect of AI regulation remains divisive for both the tech industry and the Trump Administration. Recently, White House AI advisor and a16z general partner Sriram Krishnan dismissed the possibility of an AI regulator within the executive branch, stating, "There will not be an FDA for AI." Establishing the standards body as a self-regulatory organization like FINRA could address these concerns.

Hassabis envisions the regulator staffed by open-source representatives and technical experts from within the industry, with financial support from AI labs to retain these professionals. They could also outsource evaluations to the growing pool of AI safety groups specializing in 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 argues. "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."