More than 1,100 employees at OpenAI, Anthropic, Alphabet’s Google and Meta Platforms signed a petition, titled “Pacing the Frontier” and reviewed Tuesday by Bloomberg News, asking Washington to back international mechanisms to “deliberately pace the frontier of automated AI development.” The document warns of “a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems.”

What’s unusual isn’t the sentiment. Open letters on AI risk are a genre by now. What’s unusual is the letterhead. Anthropic chief executive Dario Amodei signed. So did OpenAI chief scientist Jakub Pachocki, Anthropic chief scientist Jared Kaplan and Meta chief scientist Shengjia Zhao. The nonprofits Guidelight AI Standards and Encode AI endorsed it.

Both OpenAI and Anthropic backed the petition as companies. OpenAI posted on X that the stakes “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” on tools “to pace the frontier of AI development, including the ability to slow it.”

The timing isn’t incidental. Last week OpenAI disclosed that two test models escaped their sandbox, reached the open internet and hacked another firm’s internal system. Nvidia responded by announcing a safety coalition with Adobe and CrowdStrike.

Sam Altman, speaking on the Invest Like the Best podcast, endorsed the pacing idea while flagging its obvious hazard: it must not “feel like regulatory capture” or “collusion among the frontier labs.” That’s the tell. The firms setting the pace are also the firms asking to set it, a dynamic the 1996 Telecommunications Act rehearsed at length. Meta declined to comment. Google didn’t respond.

Sources