1,100 AI insiders just asked Washington to slow the race

More than 1,100 current and former employees of OpenAI, Anthropic, Google, and Meta have signed a public statement asking the U.S. government to support an international effort to "deliberately pace" the development of the most capable AI systems. The petition, "Pacing the Frontier", published this week and first reported by Bloomberg via Engadget, argues the industry is locked in a race in which no single company can slow down without losing ground, and that the restraint decision therefore belongs to governments rather than labs.

Among the roughly 1,122 names compiled by TheZvi are signatories from OpenAI, Anthropic, Meta, and other AI labs. The petition is a labor-side signal from inside that engine room, not a corporate position.

The petition's own argument turns on competition rather than on any specific harm. The statement says the world's leading AI companies believe they could be close to automating AI research, and that "a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems" is on the horizon. It also names the binding constraint: "each company, and country, is under intense competitive pressure not to unilaterally slow that acceleration." Without coordination, the cheapest move for any one firm is to keep going.

OpenAI's own post on X corroborates the framing in its own words: "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." The language is conditional rather than predictive, and it places the eventual decision outside the company.

The federal government has so far taken a "haphazard" approach to AI policy, and the petition arrives after hundreds of billions of dollars in private capital has been committed to the buildout. Data center construction, energy demand, copyright disputes with publishers, and national-security concerns about model weights are all live. The White House is openly skeptical of new tech rules. Asking Washington to take the lead on "deliberately pacing" the most capable systems means asking an administration that has been hostile to AI regulation to do the opposite.

The petition is thinner than its signatory list on the mechanism. It calls for "technical and governance tools" to pace frontier progress, but does not specify what those tools would look like in practice: an international monitoring body, compute thresholds, mandatory pre-deployment review, a treaty, or something else. The signatories are naming a problem rather than proposing a policy.

The ask is for an international effort routed through Washington because no other shape matches the petition's own race-condition argument. If one country moved alone, the work would move to another, and the firms most exposed to that capital flight would resist. The petition amounts to a recognition that the boundary on AI development is going to be drawn somewhere, and that the people currently drawing it are not the right ones to do the job.

The signatories have not named a government they expect to act first, and the petition names no body, no threshold, and no enforcement mechanism. The next move belongs to the people whose names sit at the top of the list.