DeepMind CEO proposes US-led global AI safety regulator

Google DeepMind CEO Demis Hassabis is calling for the creation of a global AI safety organization led by the United States, with authority to evaluate cutting-edge models before release and coordinate industry-wide slowdowns when systems pose excessive risk.

In a blog post titled "A Framework for Frontier AI and the Dawning of a New Age," Hassabis argued that the US should spearhead the initiative based on its economic and technical position. The proposed organization would include independent experts and representatives from open-source communities, operating similarly to existing regulatory bodies like the Financial Industry Regulatory Authority.

Why it matters

As AI capabilities accelerate toward what Hassabis describes as artificial general intelligence within "a few short years," the industry faces a coordination problem: no single entity can unilaterally slow development without losing competitive ground. A credible international regulator could provide the framework companies need to prioritize safety without sacrificing market position—if governments and competitors actually participate.

Building consensus behind closed doors

According to Axios, Hassabis has spent recent months quietly building support for the proposal. He has briefed the Trump administration, competing AI laboratories, and European officials on the framework. Hassabis told Axios he has received "very positive" signals from the Trump administration and hopes to launch the organization before year's end.

The proposal represents the latest attempt by industry leaders to establish governance structures for increasingly powerful AI systems. The urgency, Hassabis wrote, stems from the growing sophistication of these systems and their proximity to transformative capabilities.

Authority to intervene

The proposed watchdog would hold two key powers: evaluating frontier AI models before their public release and coordinating slowdowns across the industry when models are judged too dangerous to deploy. This authority distinguishes the proposal from advisory bodies that lack enforcement mechanisms.

Hassabis framed the current moment in historical terms, suggesting future observers will recognize this period as "standing in the foothills of the singularity—nothing less than the dawning of a new age for humanity."

The details were first reported by The Verge and Axios.