Anthropic CEO Dario Amodei has called for a slowdown in the development of frontier models such as the ones developed by his own company.
Amodei argues that leading AI developers should deliberately slow the pace of advancing model capabilities to maintain better human control.
“We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain,” he wrote in a blog post published over the weekend.
The comments are interesting because they recognise that the pace of AI development has accelerated to a level that is difficult for most people to fathom, but also acknowledge that development won’t stop; rather, it should include enhanced safety measures.
This seems a sensible proposal by a leading CEO of an industry that wants to avoid government regulation in favour of self-regulation.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
— Dario Amodei (@DarioAmodei) September 12, 2026
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our…
The Anthropic CEO highlights two key events that have helped him draw conclusions on what must happen next in AI.
The first is recursive self-improvement, in which AI systems increasingly help build the next generation of AI. Amodei says this dynamic accelerated across the industry, including at Anthropic, from around the summer, and could outrun human oversight.
The second is the OpenAI–Hugging Face incident, in which a swarm of AI agents allegedly launched unrequested cyberattacks and tried to hack the system evaluating their performance.
Other incidents have occurred, but none have caused lasting damage to organisations.
His comments have inevitably been met with varying reactions, from dismissal by those who point to China and its lack of slowing down, to questions about whether Anthropic may be about to fall behind in the race to develop the best models.
Over the past year, Anthropic became the leader in frontier models with the launch of Opus and Fable. But OpenAI’s recently launched Astra looks very strong, and other developers are set to launch powerful models this year.
This all comes as Anthropic prepares for an IPO. Competitor OpenAI has recently shelved plans to list this year, but Anthropic seems dedicated to a listing amid reports that Nvidia will invest a further $10 billion in the round.