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…
— from a petition “Pacing the Frontier”, signed by 1,319 key employees from America’s leading AI companies and published on July 28, 2026.
I’ve been writing in this editorial series about local elected officials here in Pagosa Springs who have approved — or are considering the approval of — moratoriums on data centers within Archuleta County, and possibly permanent bans in the future. The reasons given have tended to be practical — negative impacts on water and electrical resources, for example — but behind those practical concerns may linger larger political considerations.
The 1,319 signatories who added their names to the ‘Pacing the Frontier’ petition this month were not specifically addressing the problems and promises of data center development per se.
Their concern was the rapid and seemingly uncontrolled development of “frontier AI”.
That is to say, “automated AI”… AI algorithms have reach a point where AI begins to direct its own development and growth, something the industry calls “Recursive Self-Improvement.” Anthropic has shared that its ‘Claude’ models are quickly approaching that threshold.
The open letter, addressed to the U.S. government, is signed by top officials from Anthropic, OpenAI, Meta, and Google, including two Anthropic cofounders, Jack Clark and Jared Kaplan; OpenAI’s chief scientist Jakub Pachocki; Meta’s chief scientist Shengjia Zhao; and Anca Dragan, who leads AI safety and alignment at Google DeepMind.
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.
The people signing the letter are in fact the very people who are creating the automated computer systems capable of evolving beyond their own creators’ ability to control them.
The letter is a cry for help.
Obviously, these individual scientists and engineers could feasibly — on their own — put a stop to this potentially dangerous evolution. They could, themselves, pull the plug on continued development. To judge by the letter, things have not yet spiraled out of control.
But every AI company and every tech-savvy nation is seemingly locked in a race toward a future they all fear, because the other guy — and the other guy’s AI system — might get there first.
From one of the signatories, Dawn Song, Vice President of AI Research at Meta:
Frontier AI capabilities are advancing rapidly, and the pace of progress is itself accelerating.
We see this directly in our evaluation work. CyberGym and ExploitGym show that frontier AI agents are now capable of discovering and exploiting real-world software vulnerabilities, which without appropriate safeguards, could enable cyberattacks at scale…
Deliberate pacing is a heavy-handed and potentially extreme measure, and we may never need it. But if we do, it cannot be safely invented in the middle of a crisis…
…Such a capability must be international and measurable: international because no single lab or country can pace a frontier everyone else is racing toward; measurable because effective governance requires rigorous, transparent, and verifiable evaluations of frontier AI systems…
…Building the capacity is not the same as deciding when, or whether, to use it. The latter requires great caution, rigorous evidence, and broad international consensus. But without tools in place, there will be no choice at all…
If we accept Ms. Song’s analysis as accurate, one of the biggest threats seems to be AI systems capable of maliciously assuming control of other computer systems.
The White House published ‘Executive Order 14409’ on June 2, 2026, suggesting “voluntary” cooperation between the federal government and the AI industry that would somehow allow America to remain at the forefront of (uncontrolled?) AI development while simultaneously ensuring that government and business systems are protected from attack. From that Executive Order:
As these [AI] capabilities evolve, my Administration will continue to work closely with industry to ensure that the best and most secure technology is deployed rapidly to confront any and all threats to our country. We will continue to lead an America First cybersecurity effort that enhances both our national security and our global AI dominance…
Two months later, 1,319 industry leaders are demanding a different approach:
International cooperation to oversee the “pacing” of “frontier AI”.
Their letter comes in the wake of a recent incident in which AI systems from OpenAI “broke out of their testing environment and hacked into the tech startup Hugging Face.”
According to my research, the U.S. already has around 5,000 data centers full of hard-working computer servers operating with minimal supervision. We can hardly doubt that some of these millions of servers are — right now — hard at work discovering new ways to commit cyber-crimes and stage attacks on governments and businesses.
Local governments, state governments, national governments.
Local businesses, national businesses, international businesses.
The current federal administration seems to believe that we can have our America-First cake and eat it too. But it appears very likely that we are, ourselves, about to be eaten by an out-of-control cake.
In such an environment, the only governments that seem willing to try and slow down the “pacing” are local and state governments, by preventing the unchecked growth of data centers.