Google DeepMind chief Demis Hassabis calls for U.S. to spearhead AI standards body

Google DeepMind chief Demis Hassabis urged the U.S. to create and lead a standards body to evaluate frontier AI models before deployment, citing rising national security risks such as cybersecurity, nuclear, and biological threats. He proposed a federally overseen public-private partnership, similar to FINRA, with substantial industry funding and technical experts to test models up to 30 days before release and require mandatory review for U.S. deployment once effective. The article highlights growing debate over AI regulation as the U.S.-China AI race intensifies and as leading companies like Anthropic and OpenAI face increasing government scrutiny and export/rollout restrictions. The market relevance is primarily for AI-related mega-cap tech names and the broader sector, as tighter standards could affect model rollout timelines, compliance costs, and competitive positioning.