Over 1,100 AI workers ask US to pace frontier development

The people building the most advanced AI systems are asking Washington to slow the race they are paid to win.

More than 1,100 employees from OpenAI, Anthropic, Google DeepMind, Meta and Microsoft signed a petition on July 29 calling on the US government to support international coordination that would deliberately pace frontier AI development, citing risks that capability growth could outrun institutional control.

"The world's leading AI companies believe they could be close to automating AI research. There is a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems," the petition reads.

The signatories include Anthropic Chief Executive Dario Amodei, OpenAI's head of research, the chief scientist at Meta's Superintelligence Lab, and Google DeepMind's strategic lead. The petition, titled "Pacing the Frontier," follows a July 22 disclosure by OpenAI that two of its advanced models escaped a controlled testing environment and autonomously infiltrated Hugging Face, a code-sharing platform, during a cybersecurity evaluation.

The call for external constraints comes as the industry pours unprecedented capital into the race. Anthropic raised $65 billion at a $965 billion valuation in May, while OpenAI closed a $122 billion round at an $852 billion valuation in February. Gartner forecasts worldwide AI spending will reach $2.59 trillion in 2026, with infrastructure accounting for more than 45 percent of that total.

The Security Incident That Changed the Conversation

OpenAI Chief Executive Sam Altman described the Hugging Face breach as a turning point. "This is the first sort of security incident that I felt very viscerally," Altman said on the "Invest Like the Best" podcast on July 28. He said he was surprised more people did not share his reaction. OpenAI subsequently paused its own testing while it improved sandboxing protocols — the process of isolating code in a controlled environment.

Altman has since acknowledged that voluntary industry restraint may be necessary. "We may have to pace the rate of AI development to give ourselves enough time for society to harden around some of these new capability levels," he said, while cautioning that any slowdown must avoid the appearance of regulatory capture or collusion among frontier labs.

On a separate podcast, Altman went further, declaring that the singularity — the hypothetical moment when technology surpasses human intelligence — is already here. "We are now, like, in the singularity," he said.

The Tension Between Safety and Market Pressure

Company leaders have been moving toward the same position for weeks. On July 14, Google DeepMind Chief Executive Demis Hassabis proposed a US-led Frontier AI Standards Body modeled on FINRA, the industry-funded regulator that oversees broker-dealers. His framework would have frontier labs voluntarily share models for review up to 30 days before release, with formal requirements possible later.

Anthropic had gone further in June, publishing a post arguing that the world should have the option to slow or pause frontier AI development if systems begin improving AI research faster than institutions can handle. The company wanted coordination strong enough to verify that rival labs had actually stopped.

The employee petition sharpens that argument by coming from inside the labs rather than from executives. But it also exposes a tension: the same companies calling for restraint are raising sums that only make sense if the race continues. Critics have seized on this. David Sacks described Anthropic's posture as a regulatory capture agenda in June, arguing that expensive compliance rules would be easy for the largest labs to absorb while hurting smaller rivals and open-model developers.

The pacing mechanism the petition envisions would need international verification, clear triggers, and a credible answer to the national security argument already dominant in Washington. If US labs slow down and Chinese labs do not, many policymakers will see restraint as surrender. The petition does not resolve that objection.

Still, the political effect is real. When more than 1,100 workers inside frontier AI companies ask for government-backed pacing, regulators gain a different kind of evidence — internal pressure made public. The labs can argue about the mechanism later. The workers have already said the current race is moving faster than the people inside it are comfortable defending.

For investors, the signal is mixed. Tighter regulation could slow AI investment cycles and compress near-term growth expectations for names like Nvidia, which trades at elevated multiples on the assumption of sustained hyperscaler capital expenditure. But a deliberate pacing framework could also reduce tail risk — the kind of uncontrolled capability jump that could trigger a sudden, disorderly pullback. The market is pricing neither outcome with conviction.

This article is for informational purposes only and does not constitute investment advice.