DeepMind chief Demis Hassabis has called for the creation of a US‑led body to test frontier AI models, arguing that such an initiative is essential to manage the risks posed by the most advanced artificial intelligence systems.

He warned that society currently enjoys a precious window to prepare for technology that is advancing at historic speed, and that decisive action now could shape the trajectory of AI development for decades to come.

The appeal comes amid growing concerns among policymakers and researchers that unchecked AI progress could outpace existing regulatory frameworks, potentially leading to unintended consequences.

A Nobel laureate echoed the sentiment, stressing that proactive governance is needed to harness AI’s benefits while safeguarding against existential threats.

Hassabis emphasized that the proposed US‑led entity should bring together government, industry, and academia to establish rigorous testing standards, transparency requirements, and safety benchmarks for frontier models before they are deployed widely.