DeepMind CEO urges US to create AI watchdog with pause power

The chief executive of Google's AI division has called for the United States to establish a new regulatory body with the authority to screen the world's most advanced artificial intelligence models and potentially order a slowdown in development if risks are deemed too severe. Demis Hassabis, CEO of Google DeepMind, detailed the proposal in a public statement on July 14, 2026, urging that such a watchdog be operational within months, ideally before the end of the year.

Hassabis argued that the accelerating capabilities of frontier AI systems, which could approach artificial general intelligence within a few years, present urgent and growing risks. He specifically warned that serious biological, nuclear, and cybersecurity threats could emerge from advanced models in the near future, necessitating a systematic oversight framework beyond current self-governance by tech companies.

The proposed body is modeled on the Financial Industry Regulatory Authority (FINRA), the private, industry-funded organization that oversees U.S. broker-dealers under the authority of the Securities and Exchange Commission. Under Hassabis's plan, the AI standards body would be a public-private partnership...