CEO Urges Global Speed Limit On AI Development

The chief executive officer of Anthropic, Dario Amodei, has called for a global speed limit on the development of Artificial Intelligence (AI), warning that its rapid advancement capabilities could outpace efforts to make the technology safe.

Amodei made the call in a September 2026 essay titled “We Must Pace the Frontier.”

In the essay, he said governments and leading AI companies must slow the rate of capability development while continuing to pursue the benefits of the technology.

According to him, AI has the potential to transform human life by accelerating economic growth, improving healthcare and helping to address major diseases, but its growing capabilities also present serious risks.

“I believe that AI could cure most major diseases in the next 5-10 years, greatly accelerate economic growth rates, create a world of abundance and empowerment, and usher in a renaissance of democracy and freedom,” Amodei said.

However, he warned that the technology could also create risks including loss of control over AI systems, cyberattacks, bioterrorism and widespread economic disruption.

Amodei also said his position had changed in recent months because AI development was accelerating at a rate that could make it increasingly difficult for researchers and regulators to keep safety measures in step with technological progress.

“But like many technologies before it, AI brings risks, and because it is such a powerful technology, these risks are serious. We must slow the pace at which we improve the capabilities of AI models,” he said.

Amodei further said slowing AI development would provide companies with more time to improve operational security, AI alignment, interpretability, testing and evaluation, noting that AI systems had become increasingly capable of producing unexpected or undesirable behaviour, making conventional safety testing more difficult.

“More intelligent models are more capable of deceiving tests, and thus may appear aligned while having serious problems that go undetected,” he said.

He also called for stronger regulation of frontier AI companies, including permanent third-party oversight and auditing, adding that AI companies should work with governments to establish rules that ensure the growth of AI capabilities remains balanced with safety.

He called for restrictions on the sale of powerful AI chips and semiconductor manufacturing equipment to China, stronger measures against chip smuggling, action against unauthorised AI model distillation and improved security to prevent the theft of AI model weights.