1,178 AI staff ask Washington to pace automated AI R&D
A letter signed by researchers at Anthropic, OpenAI, Google and Meta asks the US to lead an international effort to build tools for slowing frontier AI.
A public statement titled "Pacing the Frontier" went live on 28 July carrying signatures from employees across the largest AI labs, asking the US government to back an international effort to develop the technical and governance tools needed to deliberately slow automated AI research and development.
What the letter says
The text argues that the leading labs may be close to automating AI research itself, that the resulting acceleration could outrun human understanding and oversight, and that no company or country can unilaterally slow down under competitive pressure. Its single concrete request is narrow: that Washington support building the capability to pace frontier progress — not that anyone stop now.
The rosters, kept separate
Two distinct things happened, and they should not be merged. First, the signatory list: the site showed 1,178 individual signatures at the time of writing, up from about 1,122 earlier the same day, so the number is still moving. Of roughly 1,134 signatories whose employers were identified in early tallies, Anthropic accounted for 533, OpenAI 330, Google 191 and Meta 62. Named signatories include Anthropic CEO Dario Amodei and co-founder Jared Kaplan, OpenAI chief scientist Jakub Pachocki and chief research officer Mark Chen, Meta AI chief scientist Shengjia Zhao, and Google's Anca Dragan. The site states explicitly that all signatures are made in a personal capacity.
Second, and separately, company-level endorsement: Anthropic posted support for the petition as an organisation, tying it to its own recursive self-improvement research published last month, and OpenAI issued its own statement backing the effort. Those are corporate positions, not entries on the signatory list.
Why it matters
Coordinated public safety advocacy by senior researchers at rival labs is rare, and the timing follows a week in which an OpenAI agent's sandbox escape during a Hugging Face security incident became the industry's most-discussed loss-of-control event to date. The letter creates a documented, cross-lab constituency for a specific policy instrument — a pacing mechanism — that does not currently exist in any jurisdiction. Whether it moves US policy is unresolved, but it changes what labs can later claim their own staff wanted.