DeepMind's Demis Hassabis has proposed a new international watchdog to test and review advanced AI models before public release, with independent experts given up to 30 days to examine each model and financed by leading AI companies.
OpenAI and Elon Musk have already backed the idea, and Hassabis plans to meet US policymakers next week to develop it into a workable system.
The proposal was prompted by concerns over Anthropic's Mythos model's cybersecurity capabilities and fears that increasingly powerful AI could pose biosecurity risks within the next couple of years.
What happened
DeepMind CEO Demis Hassabis proposed creating an international watchdog organization staffed by independent technical experts, financed by leading AI companies, with up to 30 days to examine new advanced AI models before they reach the public. OpenAI and Elon Musk quickly voiced support, and Hassabis is expected to meet US policymakers in Washington next week to advance the framework.
Why it matters
The proposal emerged after concerns about advanced cybersecurity capabilities in Anthropic's Mythos model, which Hassabis called a warning for society. He argued that increasingly powerful AI systems could create biosecurity risks within the next couple of years, and that the current ad hoc process for handling model releases may not be sustainable as AI advances.
What to watch
The plan faces political obstacles because the US, EU, and China have taken different approaches to AI regulation, and Congress has yet to approve meaningful federal legislation. Hassabis acknowledged that publishing the proposal is only the first step; whether formal international oversight can actually be established remains uncertain.
Ask the AI about this article →
DeepMind's watchdog proposal reflects growing tension between the pace of AI development and the perceived need for oversight. Hassabis framed the idea as a response to the limitations of ad hoc governance: while Google DeepMind has been discussing its newest models with government officials and AI security institutes as systems reach certain scores and benchmarks, he argued this approach may not scale as capabilities advance. The model's timing is significant—it comes as Anthropic and OpenAI have already delayed broad releases of their latest models under pressure from the Trump administration, signaling that government scrutiny is already reshaping release strategies.
However, the proposal must navigate fragmented global regulation. The US, EU, and China have taken different approaches to AI governance, and Congress has yet to pass meaningful federal legislation on the technology. These jurisdictional differences make the creation of a truly international watchdog body difficult, though Hassabis expressed belief that sufficient momentum may now exist to establish a formal structure. For AI developers, the proposal underscores that testing requirements and government scrutiny are becoming material business considerations; for policymakers, it offers a concrete model for coordinating oversight without centralizing power in any single government.
For example, today's edition would include:
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. Q&As are published on this page for other readers too.
