Google's Frontier AI Standards Proposal Faces Expert Scrutiny

Demis Hassabis, CEO of Google DeepMind, released an essay framing artificial general intelligence (AGI) as a transformative technology comparable to the discovery of electricity. He argued its impact would be unprecedented—potentially 10 times greater than the Industrial Revolution but occurring at 10 times the speed. To address safety concerns, Hassabis proposed establishing a federal regulatory body called the Frontier AI Standards Body, modeled after financial regulators like FINRA. The agency would evaluate advanced AI systems prior to release, enforce safety standards, and coordinate development slowdowns if risks escalate. While some industry leaders, including Anthropic's Jack Clark, expressed support, prominent critics questioned the proposal's adequacy. Experts like Nate Soares and Aaron Scher pointed out that risk assessments among researchers span from 5% to 90% existential risk—a range Hassabis glossed over. They also noted the framework fails to address risks from internal model deployment, which accounts for much of the danger. The discussion grew more contentious following Alex Turner's resignation from Google DeepMind. Turner quit in protest over the company's agreement with the U.S. Department of Defense to provide unrestricted military access to its models. Turner had previously attempted to block this deal, citing DeepMind's founding principles—commitments Hassabis maintained remained intact despite contrary evidence.