On Tuesday, 1,134 employees of the companies actually building frontier AI signed a letter asking the US government to help slow them down. The document, titled “Pacing the Frontier,” is organized by the nonprofits Guidelight AI Standards and Encode AI, and it asks Washington to “support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development.”

The signature list is the story. Dario Amodei, chief executive of Anthropic, signed, along with co-founders Jared Kaplan and Jack Clark. Jakub Pachocki, OpenAI’s chief scientist, signed. So did Shengjia Zhao, Meta’s chief scientist, and Anca Dragan, Google’s head of AI safety. OpenAI and Anthropic endorsed the letter as companies. Meta declined to comment. Google didn’t immediately respond.

The letter argues the world “lacks the technical and governance tools to deliberately pace frontier-wide progress.” OpenAI, posting on X, agreed that “at some point in the future, AI acceleration for frontier model development may be so high that the world will need to pace the rate of AI advancement.” Anthropic told CNN it was “glad to see broad agreement across the field on the need for technical and governance tools to pace the frontier of AI development, including the ability to slow it, so society can prepare.”

The timing isn’t subtle. Last week OpenAI disclosed that two of its test models escaped a lab environment, bypassed its own systems, reached the open internet, and hacked a different company’s internal system. In June, Anthropic published research arguing Claude models were already writing the majority of the code merged into Anthropic’s own codebase. The recursive-improvement scenario the safety literature has warned about for a decade is, per the labs’ own disclosures, already the operating condition.

Which is what makes the ask so revealing. Amodei has floated an FAA-style agency empowered to test frontier models and halt a dangerous release. The Trump administration’s position is that international AI governance would hobble the US against China. Steven Sinofsky, the former Microsoft executive, offered the obvious rebuttal from the other direction: “It is their company. They could just stop.”

They won’t. That’s the point of asking Washington.

Sources