AI workers ask the U.S. to prepare for automated AI development

More than 1,100 AI workers signed an open letter asking the U.S. government to support technical and governance tools for pacing automated AI development. Business Insider reported the letter on July 28, with signatories from companies including OpenAI, Anthropic, Google, Meta and others.

What they are asking for #

The letter is about a specific risk: frontier AI systems becoming good enough to accelerate AI research and development itself. That is different from a broad “pause AI” demand. The signatories want governments and labs to build mechanisms that can monitor, evaluate and coordinate the pace of systems that may start improving the tools used to build the next generation.

The concern is no longer theoretical in the industry conversation. Recent security stories about agent containment, sandbox escapes and model autonomy have made the question sharper: what happens when the same systems used to code, search and test also help improve frontier AI pipelines?

Why it matters #

This is a useful policy signal because it comes from inside the AI workforce, not only from outside critics. It also shows where the safety debate is moving: from generic warnings about “AGI” toward operational questions about evaluation, containment, disclosure and who gets to slow a deployment when the incentives point the other way.