DeepMind CEO Calls for Independent AI Standards Body

DeepMind CEO Demis Hassabis has put forward a bold proposal to create an independent standards body for regulating frontier artificial intelligence. In a recent interview, Hassabis suggested modeling this new organization after FINRA, the Financial Industry Regulatory Authority, which oversees securities firms in the United States. The proposed body would be responsible for testing frontier AI models before they are released to the public. It would develop best practices for safe deployment, establish benchmarks for performance and safety, and hold companies accountable for violations. Hassabis argues that the current patchwork of voluntary commitments and internal safety reviews is insufficient given the rapid pace of AI advancement. “We need an independent, expert-led organization that can evaluate these models objectively,” Hassabis said. “The stakes are too high to leave safety entirely in the hands of individual companies.” The body would have the authority to conduct audits, require model modifications, and even halt releases if safety standards are not met. This call comes at a time when concerns about AI safety are growing. Recent incidents involving jailbroken models, biased outputs, and potential misuse have highlighted the need for robust governance frameworks. Hassabis’s proposal is notable because it comes from within the industry, signaling that even AI leaders recognize the limits of self-regulation. While the idea has garnered support from some researchers and policymakers, it also raises questions about implementation. Who would fund such a body? How would it keep pace with rapidly evolving technology? And would companies comply voluntarily? Despite these challenges, Hassabis’s proposal represents a significant step toward a more structured and accountable AI ecosystem.