More than 1,100 workers from leading AI companies today published an open letter asking the U.S. government to help create tools that can deliberately slow automated AI model development, citing a growing risk that such systems could advance faster than humans can understand or control them.
The document gathered 1,134 signatures from employees at firms building frontier AI models, including Anthropic, OpenAI, Google, and Meta. Staffers at venture-backed labs like Thinking Machines and London-based Inherent Laboratories also signed. Notable names include Anthropic CEO Dario Amodei and several of his co-founders, along with the chief scientists of OpenAI, Meta's Meta AI unit, and Thinking Machines.
What the letter says
"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 reads. The signatories argue that the industry currently lacks the means to "deliberately pace" automated AI research, and they point to the need for new technical and governance mechanisms.
"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," the letter states. The appeal does not propose specific policies but frames the problem as a tools gap that only coordinated government action can close.
Where automation is already accelerating research
The letter comes as several AI labs push to automate their own model development pipelines. The letter focuses on the use of frontier models like GPT and Claude to automate tasks that human researchers once handled, an area covered under Generative AI and LLM. Last month, OpenAI disclosed that its GPT-5.5 large language model helped optimize the graphics card cluster it runs on, boosting token generation speeds by more than 20%. OpenAI has since released a more capable GPT-5.6 series with sharper coding skills.
Anthropic similarly used its Claude model to speed up AI safety research, building a Claude-powered agent to study weak-to-strong supervision, a technique the company believes could be key to supervising future AI systems. These efforts highlight how AI is already being used to accelerate its own improvement, a trend that falls squarely within AI Agents & Automation.
Broader calls for guardrails
The open letter is the latest in a series of recent policy pushes from within the AI community. On Friday, Nvidia and other large tech firms urged U.S. policymakers not to ban open-source AI models. A few days earlier, a group of economists called for stronger AI regulation. Earlier this month, Google DeepMind CEO Demis Hassabis proposed a new standards body to evaluate large language models before they are broadly released.
Why this matters for IT and Development
For developers and IT professionals, the accelerating automation of AI research shifts the ground under their daily work. When models like GPT-5.5 can optimize their own infrastructure and Claude agents can drive safety research, the pace of model releases and the security demands of the software supply chain both intensify. The tools and governance the letter asks for could directly affect how quickly new AI capabilities land in production environments, how much testing they receive, and what safety benchmarks teams must meet before deployment.
Your membership also unlocks:
