DeepMind CEO calls for an independent standards body to regulate frontier AI

Editor's take

DeepMind CEO Demis Hassabis has proposed the creation of an independent body to establish and enforce standards for the development and deployment of advanced AI systems. This initiative emerges from growing concerns about the potential risks associated with increasingly capable frontier models, such as those developed by Google's own DeepMind, OpenAI, and Anthropic.

The call for a regulatory framework, drawing parallels to financial industry oversight like FINRA, highlights a significant shift in the AI discourse from pure innovation to risk management. It acknowledges that unchecked development of powerful AI could outpace society's ability to control it, impacting everything from scientific discovery to national security. This proposed body would aim to provide a much-needed layer of accountability and safety assurance for the public and policymakers alike.

Future developments will likely focus on the practical implementation of such a body: who would fund it, which organizations would participate, and what specific testing and certification criteria would be established for models like GPT-5 or Google's Gemini. The willingness of major AI labs to cede some control to an external entity will be a crucial indicator of the seriousness with which safety is being prioritized.