Google DeepMind CEO Demis Hassabis is proposing a new, powerful regulatory body to oversee the world's most advanced AI models. In a recent interview, Hassabis outlined a vision for a US-government-led agency that would be funded by the AI industry and staffed by elite technical experts.
This watchdog would have authority to screen AI systems before their release and could coordinate an industry-wide slowdown if safety risks emerge. The proposal calls for mandatory pre-release testing to prove models are safe before public deployment.
Hassabis has previously warned about 'race conditions' in AI development, where competitive pressure pushes labs to ship products faster than they can verify safety. He has suggested Artificial General Intelligence could arrive by 2029, adding urgency to the regulatory discussion.
The proposal carries significant weight as it comes from the head of a leading AI lab, effectively asking for constraints on his own industry. However, some observers note such a framework could favor well-resourced incumbents like Google DeepMind, potentially disadvantaging smaller competitors and open-source development.
While Hassabis advocates for global cooperation, his proposal anchors regulation in American governance, betting on the US's institutional credibility and technical talent to lead the effort.