OpenAI and Anthropic Both Call for a Global AI Watchdog as Both Prep for IPOs

Rivals Find Rare Common Ground

In a striking moment of alignment between two of the fiercest rivals in artificial intelligence, both OpenAI and Anthropic are now publicly backing the idea of an international organisation empowered to oversee, and if necessary slow, the development of the most powerful AI systems.

OpenAI CEO Sam Altman and chief scientist Jakub Pachocki published a blog post on Monday calling for the formation of an international body to govern frontier AI development, including the ability to slow progress "when needed" so that safety and societal resilience can keep pace. The call formed part of a wider vision statement announcing what Altman described as OpenAI's "third phase" — a shift away from pure research breakthroughs and product launches towards making advanced AI universally accessible and safe.

The Timing Is No Accident

The announcement landed the very same day OpenAI confirmed it had confidentially filed for an initial public offering with the SEC, with the company valued at more than $850 billion. It also followed closely on the heels of a report from rival Anthropic, published on 4th June, urging AI labs to develop a coordinated and verifiable framework for pausing development should advanced systems begin improving themselves faster than society can control.

"We believe it would be good for the world to have the option to slow or temporarily pause frontier AI development to enable societal structures and alignment research to keep pace," Anthropic wrote. The company warned that "recursive self-improvement" — where AI systems independently design, build and train their successors — could arrive sooner than many organisations are ready for.

From Competition to Coordination

The convergence is remarkable given the intense commercial rivalry between the two firms. Both are preparing for public offerings, with Anthropic having filed its own S-1 in early June. Yet both now endorse the notion that an external authority should be able to throttle the most powerful systems.

The push is not entirely new. Altman first floated an "IAEA for superintelligence" back in 2023, but this latest statement is more concrete. Anthropic stressed that any meaningful pause would require agreement from multiple leading labs across nations, particularly the United States and China, alongside robust verification protocols.

Rhetoric or Reality?

The central question is whether these proposals amount to more than words. Neither company has committed to unilateral limits on its own development, and Anthropic acknowledged that if only one firm paused, competitors would simply surge ahead. The challenge now falls to governments: will they accept the invitation, and will the labs hand over genuine external authority? With AI pioneer Yoshua Bengio recently warning the world is building systems "we don't know how to control," the pressure for coordination has rarely been higher.