Claim◆Article
Over 1,000 employees from frontier AI labs have cosigned a statement requesting the U.S. government support an international effort to develop tools needed to deliberately pace the frontier of automated AI development, due to the risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems.
A letter signed by over 1,000 frontier AI lab employees—from OpenAI, Anthropic, Google DeepMind, Meta, and others—calls on the U.S. government to support international technical and governance tools to deliberately slow frontier AI development if needed, citing risks from recursive self-improvement. ✦ AI generated
1,171 employees of frontier AI companies · Latent Space · 2026-07-29 · original ↗
"AI could help create a dramatically better future, but that outcome is not guaranteed. The world's leading AI companies believe they could be close to automating AI research. 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. To realize AI's potential, industry, government, and society at large may need the option to buy time to address emerging risks, develop security measures, and strengthen oversight. But each company—and country—is under intense competitive pressure not to unilaterally slow that acceleration. And today, the world lacks the technical and governance tools to deliberately pace frontier-wide progress. Building on work already underway to monitor frontier model releases: 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." - 1,171 employees of frontier AI companies
Read full article ↗excerpt · fair-use quotation
- ·Signatories from OpenAI, Anthropic, Google DeepMind, Meta
- ·Warn of recursive self-improvement accelerating beyond control
- ·Urge US-backed international tools to pace frontier development
- ·No company can unilaterally slow without losing competitive edge
- ·Neither technical nor governance pacing tools exist yet
- ·Letter asks US to lead international effort to build them
Around this claim
This moment responds to
rebuts → An FAA-style regulator for AI would be an extremely slow, permission-based system — modeled on aircraft type certification that takes 5 to 9 years — and applying that logic to AI, which currently ships new versions every couple of months, would cause the US to lose the AI race to China.David Sacks · All-In Podcastsupports → The US government should support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development.the Pacing the Frontier letter signatories · Simon Willison's Weblogrebuts → It is far too early to impose heavy regulation on American AI because regulators don't yet understand what actually needs to be regulated, much like early automobile regulation took decades to figure out.Kevin Weil · BG2 Podsupports → The US government should stand up a federally overseen, FINRA-like Standards Body to test frontier AI systems for dangerous capabilities, beginning with voluntary pre-release review before moving toward formal regulation.Demis Hassabis · Import AIsupports → The US government should create a FINRA-style Standards Body to voluntarily test frontier AI systems for dangerous capabilities, moving toward formal regulation once the process proves effective.Demis Hassabis · Import AIprovides context → The U.S. government asked OpenAI to stagger GPT-5.6 preview access customer-by-customer, signaling an emerging de facto government review regime for frontier model launches.The Information (via AI Twitter Recap) · Latent Space