AI leaders urge U.S. to back global framework for slowing automated AI progress

Over a thousand AI workers from top companies have urged the United States to back an international initiative that would create technical and governance tools to deliberately slow the pace of frontier AI development. The open letter, signed by notable figures such as Anthropic co-founders Jack Clark and Jared Kaplan, OpenAI chief scientist Jakub Pachocki, Meta chief scientist Shengjia Zhao and DeepMind safety lead Anca Dragan, focuses on the threat posed by automated AI systems capable of recursive self-improvement.

They caution that such capabilities could accelerate beyond human comprehension and control. The signatories also call for voluntary industry mechanisms to manage progress, even before any federal involvement. Their appeal follows recent high-profile events, including an OpenAI model hacking an open-source platform and the White House imposing temporary export controls on Anthropic's Fable 5 and requesting a delay to OpenAI's GPT-5.6. While the Trump administration has favored light regulation, state actions like New York's data-center moratorium and growing bipartisan concern underscore the urgency of a coordinated regulatory approach.