Demis Hassabis urges U.S.-led AI standards body
AI safety standards body could help the U.S. assess frontier model risks Review AGI systems for cyber, bio, and nuclear threats before deployment
Google DeepMind chief Demis Hassabis is calling for the United States to lead a new standards body that would review frontier AI models for national security risks.
In comments posted on X, Hassabis said rapid advances in artificial general intelligence require urgent action, citing cybersecurity concerns and possible future bio and nuclear risks. He suggested a federally overseen publicprivate partnership, with substantial industry funding, to test advanced models before deployment.
The proposal comes as debate grows over how to regulate leading AI systems and as competition between the U.S. and China intensifies. Hassabis said such a body could evaluate safety guardrails, watermark AIgenerated images, and help establish common technical standards.