Hassabis urges U.S.-led AI standards body
Demis Hassabis, head of Google DeepMind, called for a U.S.-led standards body to evaluate frontier AI models for national security risks including cybersecurity and biological threats. In a public post he proposed a federally overseen public-private partnership—modeled on organizations like FINRA—with independent technical experts, open-source representation, and substantial industry-funded resources to test models. Frontier labs would initially share models voluntarily up to 30 days before release, with mandatory review for U.S. deployment once the body proved effective. The call follows similar proposals from other industry leaders at a recent G7 meeting and amid heightened U.S.-China competition in advanced AI model development.
- •Demis Hassabis (Google DeepMind chief) called for a U.S.-led standards body to assess frontier AI models for national security risks.
- •Hassabis proposed a federally overseen public-private partnership or self-regulatory organisation modeled on FINRA, with independent technical experts and open-source representatives.
- •Frontier labs would initially voluntarily share models with the proposed body up to 30 days before release, becoming mandatory for U.S. deployment after effectiveness is demonstrated.
