Google DeepMind CEO Demis Hassabis has stated that artificial general intelligence is only a few years away and proposes a new U.S. standards body to test frontier AI models before release. He described AGI as potentially bigger than electricity or fire in its impact on humanity.
Speaking about the rapid progress in AI development, Hassabis emphasized the need for proactive governance of advanced AI systems. He called for the establishment of a standards body modeled on the Financial Industry Regulatory Authority, with a board that includes independent experts and open-source representatives.
Hassabis proposed that this organization would be responsible for developing assessment protocols and working with federal agencies and U.S. National Labs to conduct testing in areas relevant to national security. AI vendors would be encouraged to adopt best practices including publishing model cards, maintaining cybersecurity standards, and providing resources for safety research.
The DeepMind CEO has been working behind the scenes to build support for this proposal among the Trump administration, other AI companies, and European policymakers, with the goal of having the organization operational before the end of 2026.
Hassabis has previously participated in U.S. government initiatives evaluating AI safety, working alongside Microsoft and SpaceXAI with the Center for AI Standards and Innovation. He argues that as AI capabilities advance rapidly, waiting for traditional regulatory processes would be too slow.
The call for self-regulation comes amid growing concerns about AI safety from both industry leaders and government officials. Major AI companies have been increasingly vocal about the need for coordinated governance of frontier AI systems.
This article was adapted from Decrypt. Read the original here.
