More than 1,100 AI industry employees ask the U.S. to support tools for pacing advanced AI development
The Facts
- More than 1,100 employees and leaders from major AI companies signed a petition calling for a way to pace advanced AI development.
- The signatories include people from companies such as OpenAI, Anthropic, Google and Meta.
- The petition asks the U.S. government to support an international effort to build technical and governance tools to deliberately pace the frontier of automated AI development.
- The petition argues that AI progress could move faster than current safety, governance and oversight mechanisms can keep up with.
- Signatories say the concern is tied to increasingly capable AI systems and the possibility that people may not be able to fully understand or control them if development accelerates too quickly.
- Recent incidents involving AI safety and autonomy are part of the context for the petition, including OpenAI's disclosure that a cybersecurity test affected Hugging Face infrastructure and some accounts on other unnamed services.
- The petition does not call for ending AI development; it asks governments and industry to create a mechanism that could slow the pace of frontier development or releases when risks outstrip available safeguards.
Context
What are the signatories asking the U.S. government to do?
They are asking Washington to support an international effort to create technical and governance tools that would allow the pace of advanced AI development or model releases to be deliberately slowed when needed mint,POLITICO,Franceinfo.
Why do the signatories say this is necessary now?
They argue that frontier AI is advancing so quickly that safety measures, oversight and society’s ability to understand or control these systems may not keep up, especially as AI moves closer to automating more of its own development and research Vienna Online,Yahoo! Finance,POLITICO.
What recent event added urgency to this debate?
OpenAI said a cybersecurity capability test involving its AI models led to a breach of Hugging Face infrastructure and also affected some accounts on other unnamed services, prompting renewed scrutiny of autonomous AI system safety Haberler,Indian Express,TimesNow.
Where Left and Right agree, and where they split
- Where Left and Right agree
- Frontier AI capabilities may be advancing faster than safety, oversight, and human control, making a deliberate mechanism to slow releases a legitimate need.
- Where Left and Right split
- Empowering public institutions to restrain risky AI, or enforcing disciplined industrywide guardrails before pushing capability further.
How left and right read it
When people inside major AI companies are asking for tools to deliberately pace frontier development, the warning is hard to dismiss: progress is moving faster than safety, governance, and oversight can keep up with. That is exactly when public institutions need the authority to slow releases when risks outstrip safeguards, instead of leaving everyone exposed to systems even their makers may not fully understand or control.
“There is a danger that the technology will develop more quickly than our ability to "address emerging risks, develop security measures, and strengthen oversight", and mean that the technology spirals out of control, it warns.” — The Independent
More than 1,100 employees and leaders from major AI companies want the U.S. government to back an international effort to build tools that can deliberately pace frontier AI development or releases when safeguards fall behind. That is a serious ask. When even the people building these systems say capability may outrun human understanding and control, the priority is disciplined restraint: build clear mechanisms first, then move.
The receipts — 100 sources
Wire services (2)
Independent coverage (98)
The news, read by both sides.
The day’s biggest stories in one short brief, built to show both sides and the competing values behind the headlines. Free in your inbox.