OpenAI, Anthropic Staff Ask US to Pace Frontier AI Progress

The interesting thing about a petition from inside OpenAI and Anthropic asking Washington to help slow their own industry down is who is asking, not what they are asking for. Bloomberg reports that staff at the two labs are circulating a letter urging the US government to support an international effort to 'deliberately pace the frontier of automated AI development.' The document warns of 'a real risk' that AI progresses faster than people can 'understand or control.'

The timing is the most telling part. According to reporting by Rachel Metz and Shirin Ghaffary, the petition is circulating in the days after OpenAI disclosed that its tools mistakenly hacked into Hugging Face's internal systems, an incident that sits somewhere between a governance failure and a capability demonstration depending on your reading. The letter also echoes public calls from OpenAI CEO Sam Altman and Google DeepMind CEO Demis Hassabis for a new international watchdog to vet and set standards for cutting-edge models, which puts the labs' rank and file, their executive class, and their nominal rivals in rough alignment on the same policy ask.

Why this matters if you are not tracking AI policy closely: an ask from inside the labs reads very differently from an ask from outside them. It removes the political argument that oversight is inherently hostile to the industry, and it hands lawmakers a mandate they did not have to negotiate for. Whether US officials use that mandate is a separate question, but the option now exists in a way it did not a week ago.

The honest caveat is that the reporting is thin on specifics. Bloomberg does not disclose how many people have signed, whether Google DeepMind staff are among them, or what 'deliberately pace' would look like in practice, whether that means mandatory pre-release evaluation windows, staged rollouts, a genuine pause on certain training runs, or something considerably looser. Take the demand as directional, not detailed.

For anyone building on top of frontier APIs, the thing worth watching is not the letter itself but where its logic first shows up in the real world. That is usually not statute; it is procurement questionnaires, enterprise contracts, and staged-release policies that quietly acquire 'pace the frontier' language months before any lawmaker gets around to acting.