DeepMind's Hassabis Pushes for Global AI Standards Body

Google DeepMind CEO Demis Hassabis is calling for a new AI standards body to regulate frontier models, proposing the U.S. lead the effort. He has already held talks with the White House, European officials, and rival AI labs, aiming to launch the organization by year's end. Modeled after the Financial Industry Regulatory Authority, the body would create quarterly-updated benchmarks to assess AI risks in areas like cybersecurity and biology, and detect deceptive models. Hassabis envisions a voluntary review program initially, potentially becoming mandatory for U.S. market deployment.