Google DeepMind chief calls for US-led global AI watchdog

Google DeepMind chief executive Demis Hassabis has proposed the creation of a global artificial intelligence watchdog led by the United States, aiming to have the institution operational before the end of 2026. The proposed body would comprise independent experts and representatives from open-source communities, possessing the authority to evaluate frontier models prior to release and coordinate industry-wide pauses if systems are deemed too risky.

Hassabis, who recently won the 2024 Nobel Prize in Chemistry, argues that the US is best positioned to set global standards due to its economic and technical standing. He has reportedly briefed the Trump administration and other stakeholders, receiving positive feedback, as he seeks to mitigate risks associated with the imminent arrival of artificial general intelligence.

The proposal follows statements from other tech leaders, including Anthropic cofounder Jack Clark and former Google CEO Eric Schmidt, urging world leaders to address the economic impacts of AI. Hassabis has spent months quietly building support for the initiative, which he outlined in a blog post titled “A Framework for Frontier AI and the Dawning of a New Age.”

He stated that artificial general intelligence is “probably only a few short years away” and described the current era as the “foothills of the singularity.” The proposed watchdog could resemble existing regulators like the Financial Industry Regulatory Authority, providing a structural model for oversight in a sector currently lacking comprehensive national rules in the US or global frameworks.

Hassabis recently signed a statement calling for tougher protections against AI-aided bioweapons production. The push for regulation comes as AI systems grow in sophistication and perceived risk, with the timeline for the watchdog’s establishment reported as a target rather than a confirmed deadline.