AI industry leaders briefly appeared united on supporting regulation this week, with Anthropic CEO Dario Amodei proposing a three-step plan that included third-party evaluators, industry coordination, and international agreements. However, the consensus collapsed sharply: whilst OpenAI and Anthropic pushed for government involvement in safety standards, Meta CEO Mark Zuckerberg and others blocked proposals for an independent industry regulator. This split matters because it reveals fundamental disagreement within the sector over whether—and how—to regulate AI development, even as security concerns mount.

Government support for regulation appears unlikely, with the Trump administration dismissing AI safety fears as a "hoax" and insisting that executive authority alone is sufficient oversight. OpenAI's global affairs chief argued that whilst companies must ensure their own safety practices, government should establish mandatory national standards for frontier AI, citing growing bipartisan congressional support. The debate echoes historical regulatory patterns in electricity, cars, and aeroplanes, where safety standards eventually became industry-wide norms, though early periods remained chaotic.