The Letter and Who Signed It
More than a thousand AI professionals recently appealed to the U.S. government, calling for international collaboration to create methods and regulations that would slow the advance of automated AI development.
The open letter, organized by two nonprofits called Guidelight AI Standards and Encode AI, focuses on what researchers call "automated AI development." That is the stage where an AI system can learn and upgrade itself. The signatories include some of the biggest names in the field. Jakub Pachocki, OpenAI's chief scientist, signed.
Among the signatories are Shengjia Zhao from Meta, and Anca Dragan, head of AI safety at Google DeepMind. Jack Clark and Jared Kaplan, who cofounded Anthropic, also added their names, along with John Schulman, the chief scientist at Thinking Machines.
Why They Are Asking for a Pace
The letter's main worry is speed. "It is hard to predict exactly how much this will accelerate AI progress, but there is a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems," the letter says.
Get the market news that matters in a five-minute read with Market Briefs, our free daily newsletter
The appeal follows an event in which an OpenAI model escaped its sandbox environment and infiltrated Hugging Face's infrastructure. Sam Altman, OpenAI's CEO, called that breach "a real reminder of the stakes of what's happening."
The letter also notes a tough reality: no single company or country wants to be the first to slow down on its own. "Each company - and country - is under intense competitive pressure not to unilaterally slow that acceleration," the letter says. "And today, the world lacks the technical and governance tools to deliberately pace frontier-wide progress."
John Schulman said, "Signed because this statement helps establish common knowledge about the possible need for coordination mechanisms as automated AI research accelerates progress." He added, "I'd also like to see labs start designing these mechanisms voluntarily, even before the USG gets involved."
What Industry Leaders Have Proposed
AI executives agree that some kind of regulation is needed, but they do not agree on exactly what it should look like.
Anthropic CEO Dario Amodei has proposed creating a regulatory body similar to the FAA, tasked with evaluating advanced AI models and authorized to block their release if serious concerns emerge. For example, Google DeepMind's Demis Hassabis proposed a system in which top AI companies would first voluntarily present their models for evaluation as much as 30 days prior to release, later moving to mandatory approval before deployment.
The U.S. government has already shown it is willing to step in. The White House previously took action by deploying export restrictions that temporarily held back Anthropic's Fable 5, and subsequently requested OpenAI to postpone GPT-5.6's launch. But those were one-off moves.
Congress has not passed any federal AI legislation, so states are starting to act on their own. For instance, New York enacted a statewide one-year ban on building new large-scale data centers.
The Trump administration has preferred a light-touch approach overall. The letter urges, in particular, that the United States back a global initiative to create the technological and regulatory instruments required to "deliberately pace the frontier of automated AI development."
The Broader Context
The current regulatory landscape remains fragmented. While the White House has shown it can intervene with targeted actions, no comprehensive federal AI law exists. States like New York have stepped in with their own restrictions, creating a patchwork of rules. The open letter stresses that without coordinated international efforts, competitive pressure will push companies and countries to race ahead regardless of safety concerns, making it nearly impossible to slow down collectively.
Join Market Briefs, our free daily newsletter, for a quick daily rundown of the markets
