← Back to the dossier

28 July 2026 · Employees of OpenAI, Anthropic, Google DeepMind, Meta

More than 1,300 lab employees ask for a way to slow AI down

People who build the most advanced AI signed a short statement. They ask the American government to support an international effort to build tools that can control the pace of AI progress. They do not ask for a pause now. They ask for a brake that exists before anyone needs it.

By Mara Masaeva · Updated 29 September 2026

PoliticsConfirmedMore than one independent source, or a primary document.

What happened

The statement, Pacing the Frontier, went online on 28 July 2026. It starts from what the leading companies think: they could be close to automating AI research, and nobody can predict how much faster things would then go. Every company and every country feels pressure not to slow down alone. The tools to slow down together do not exist.

The statement makes one request: We request that the U.S. government support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development.

On 29 September the site listed 1,386 signatories. They include Jakub Pachocki and Mark Chen (OpenAI), Dario Amodei, Jared Kaplan, Jack Clark and Chris Olah (Anthropic), Shane Legg (Google DeepMind), Shengjia Zhao and Dawn Song (Meta), John Schulman (Thinking Machines) and Ilya Sutskever (Safe Superintelligence). Two nonprofits, Guidelight AI Standards and Encode AI, helped organise it. Song signed as vice president of AI research at Meta. In the intelligence explosion paper she is listed with UC Berkeley.

According to TechTimes, OpenAI and Anthropic also backed the statement as companies the next day. That day the site counted 1,268 signatories.

What may follow

The statement came just over two weeks after Hugging Face. Two months later came the intelligence explosion paper, which several of the same people co-wrote. It is also the background to the lab leaders who backed a slowdown in September.

It binds nobody, and it names no threshold at which the brake should be used. Still, it is unusual for the people who would have to use the brake to sign a statement like this themselves.

My notes

A useful answer to 'if they are so worried, why don't they just stop?' They say they cannot stop alone, and they signed their names to that.

“The problem statement is real. This works only if it is done internationally, and it has to be done well: a bad implementation can make things worse.”

Ilya Sutskever · CEO of Safe Superintelligence

“I'd also like to see labs start designing these mechanisms voluntarily, even before the USG gets involved.”

John Schulman · chief scientist at Thinking Machines

Read next

Sources

  1. Pacing the Frontier: statement and signatoriesprimary · main source

    Read on 29 September 2026. The number of signatories is still changing.

  2. GAGE slowdown docket: Pacing the Frontierresearch

    Summary of the statement, the publication date and the company endorsements, which it takes from TechTimes.

  3. TechTimes: OpenAI, Anthropic Formally Back Plan to Slow AI That Writes Its Own Codepress

    Brandon Fisher, 29 July 2026. Source for the company endorsements and the count of 1,268 signatories that day.