Google DeepMind CEO Demis Hassabis has proposed a new independent standards body to help guide the release of frontier AI models. In a post titled "A Framework for Frontier AI and the Dawning of a New Age," he argued for a system that could evaluate advanced models before deployment and define shared best practices.
Under the proposal, frontier labs would first submit models voluntarily for review up to 30 days before launch. If the process proves reliable, that review could become a formal requirement for models entering the US market. The body would also work with companies after release to identify and address critical vulnerabilities.
Hassabis suggested a structure similar to FINRA, the self-regulatory organization that oversees parts of the financial industry. In his vision, the new body would be backed by the US government, funded by AI companies, and run independently by technical experts, open-source voices, and specialists in AI safety.
The idea comes as the debate over AI oversight continues to evolve across the industry. Hassabis said a technically grounded framework could support innovation while encouraging responsible development, with the flexibility to tighten standards as frontier systems become more capable. Such a model could help shape how advanced AI is introduced to society in the years ahead.
