AI
DeepMind chief Hassabis calls for US-led global AI regulator with power to halt risky models
The Google DeepMind CEO wants an international watchdog modeled on financial regulators to evaluate frontier AI systems before deployment and coordinate industry slowdowns if needed.
The Verge
Demis Hassabis, CEO of Google DeepMind, is pushing for the establishment of an international regulatory body to oversee the development of advanced AI systems and intervene if the technology poses too great a risk. The proposal marks the latest call from industry leadership for binding rules around frontier models as capabilities accelerate.
The proposed framework#
Hassabis outlined his vision in a recent blog post, suggesting the US take the lead in creating what would function similarly to existing financial regulators. The organization would bring together independent experts alongside representatives from open source development communities, granting it authority to assess frontier AI models before they reach the public and to orchestrate industry-wide deployment delays if systems appear too hazardous. According to the proposal, "the noises I've been hearing [from the Trump administration] are very positive," Hassabis told Axios.
Timing and support-building#
According to reporting by Axios, Hassabis has spent months constructing a coalition behind the initiative, engaging the Trump administration, rival AI research organizations, and European officials. He aims to have the regulatory body operational before the year's end.
Why the urgency?#
Hassabis argues that artificial general intelligence (AGI) could arrive within just a few years, making swift action imperative. His framing is sweeping—he suggested that looking back decades from now, this moment may be recognized as humanity standing "in the foothills of the singularity" as a transformative new age emerges. The lack of coherent global AI governance frameworks, combined with the absence of comprehensive national regulations even in the US, underscores the gap his proposal aims to fill.
Broader context#
The DeepMind leader's push joins a growing chorus of AI researchers and technologists calling for stronger guardrails. Hassabis himself recently joined a statement advocating stricter safeguards against AI-powered bioweapon development. Meanwhile, prominent economists and industry figures—including Anthropic cofounder Jack Clark and former Google CEO Eric Schmidt—have to address the economic disruption AI could trigger.
