OpenAI, Anthropic and Google Employees Urge US Government to Step In on the AI Race

Employees of OpenAI, Anthropic, Google, Meta, Microsoft, Mistral and Thinking Machines have signed a joint appeal to the US government. They call for slowing the development of the most powerful AI models or, at the very least, speeding up the creation of a global system of coordinated oversight for the industry. The Verge reported on the appeal in July 2026.

Who signed the appeal

The appeal was signed by employees of leading AI labs, not by the corporations in any official capacity. According to The Verge, the signatories include people from at least seven companies that normally compete fiercely for talent, compute and market share, but here they spoke with one voice.

• OpenAI and Anthropic — the developers of ChatGPT and Claude
• Google and Meta — major corporations with their own AI divisions
• Microsoft — a key partner of and investor in OpenAI
• Mistral — a French lab building open models
• Thinking Machines — a startup founded by former OpenAI employees

Such a coalition is rare: the signatories are at the same time building the very frontier models whose development they propose to slow down.

What the authors demand

The authors offer US authorities a choice of two paths: support a possible slowdown in the development of the most advanced models, or sharply accelerate work on coordinated international governance of AI. The overall message is simple: the market will not rein itself in, and the state is needed here.

"AI could help build a radically better future, but that outcome is

not guaranteed," the statement says.

The signatories stress that a good outcome will not arrive by default: a safe future has to be built through regulation and international coordination, not by hoping for the goodwill of individual companies.

Why are they afraid of AI automation?

The authors' main fear is the approach of the moment when AI starts conducting AI research on its own. The statement says explicitly that leading companies believe they are close to this threshold, and that its consequences are hard to predict.

"The world's leading AI companies believe they may be close to automating AI research.

It is hard to predict how much this will accelerate progress, but there is a real risk that the development of capabilities will spiral out of control," the appeal says.

Automating AI research means one model designs and improves the next with almost no human involvement. Such a self-reinforcing loop can push progress to a speed that neither regulators nor society can keep up with — and that is exactly what the signatories fear.

What this means

When employees of competing labs jointly ask the government to slow down their own industry, it is a strong signal. Concern about AI safety has moved beyond academic debate and reached the people who build these systems every day.

For now, the appeal is a call to action, not a law: it contains neither a concrete oversight mechanism nor a timeline. But the very fact that engineers and researchers from rival labs put their names to one text makes the conversation about slowing the AI race carry far more weight than earlier expert warnings.

Frequently asked questions

Who signed the appeal to the US government?

Employees of OpenAI, Anthropic, Google, Meta, Microsoft, Mistral, Thinking Machines and other leading AI labs. The signatures came from company employees, not from the corporations in any official capacity — a rare case of direct competitors converging on a common position.

What are the authors of the statement seeking?

They ask the US government to support a possible slowdown in the development of the most powerful models, or to speed up the creation of an international system of coordinated oversight of AI, before the industry gets close to automating its own research.

*Meta is designated an extremist organization and banned in Russia.