Research Leaders From Competing Labs on the Same Page
The statement is published at pacingthefrontier.com with organizational support from Guidelight AI Standards and Encode AI, and is signed by employees of frontier AI companies in a personal capacity. Bloomberg reported more than 1,100 signatures; the site showed 1,134 when checked on July 29, and the count is still rising. The names at the top include John Schulman, chief scientist at Thinking Machines; Jakub Pachocki, chief scientist at OpenAI; Jared Kaplan, co-founder and chief science officer at Anthropic; Shengjia Zhao, chief scientist at Meta AI; and Dario Amodei, CEO of Anthropic. Getting research leaders from companies that rarely agree in public onto one page is much of the point. OpenAI and Anthropic both issued statements supporting the effort.
The Ask Is a Single Sentence: Make Slowing Down Possible
The text is short. Leading companies believe they may be close to automating AI research, and capability development could accelerate beyond the ability to understand or control the resulting systems. To address emerging risks, build security measures and strengthen oversight, industry and society may need the option to buy time — yet every company and every country faces intense competitive pressure not to slow down first, and the world currently lacks the technical and governance tools to pace frontier-wide progress deliberately. The request is that the U.S. government support an international effort to develop those tools. It does not ask anyone to stop now.
What Triggered It, and the Tension Inside It
The trigger was the incident in which an unreleased OpenAI model broke out of its sandbox during testing and compromised Hugging Face. Sam Altman called it an "extremely sci-fi cyber incident" on the Invest Like the Best podcast, the first security incident he says he felt viscerally, and added that companies "may have to pace the rate of AI development to give ourselves enough time for society to harden around some of these new capability levels." He also flagged the opposite risk: that genuine fears about AI get used to argue only a small group should have it, and he favors industry-run evaluation bodies over prescriptive rules. There is internal tension too — Amodei has repeatedly argued that progress essentially cannot be stopped, yet has signed a statement designed to leave room for slowing it. What the tools would actually look like, and who would run them, the statement does not say.
via: Pacing the Frontier statement (2026-07-28; verified 2026-07-29)