DeepMind's Demis Hassabis Proposes U.S.-Led Global Watchdog to Screen Frontier AI Models

Google DeepMind chief executive Demis Hassabis has called for the creation of a U.S.-led global body to vet the most advanced AI systems before release and, if risks escalate, to coordinate an industry-wide slowdown in development. The proposal, laid out in a personal manifesto and detailed in an interview with Axios, envisions a standing institution with the authority to screen so-called frontier-class models and to pump the brakes on the entire industry should safety concerns mount.