Anthropic CEO Dario Amodei has called on the artificial intelligence industry to moderate the speed of model capabilities training, warning that unchecked progress could outrun human ability to safely control the technology.
“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”, Amodei wrote in an essay published on his personal website.
Amodei clarified that his proposal does not mean a complete halt to research or model training, but rather “pacing the rate of capabilities advancement so that risk prevention has time to keep up”. He warned that rapid advancements, particularly driven by systems helping build their next generation, carry serious risks, stating that “building it too fast is reckless”. As part of his proposal, Amodei committed Anthropic to granting third-party evaluators permanent, employee-level access to inspect models and verify safety measures.
The call received public backing from leaders across rival AI firms:
- Sam Altman (OpenAI CEO): Agreed with the proposal, posting on X: “I agree with Dario that we need to pace the frontier. Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We’ll have more to share soon”.
- Elon Musk (xAI / Tesla): Endorsed Amodei’s statement on X, writing simply: “Dario is right.”
While all three executives share a rare public agreement on the need for caution, the real test will be the execution.
Pacing frontier development requires AI labs to balance safety protocols with immense commercial pressure and competitive racing dynamics. Whether this consensus leads to binding industry-wide standards, independent safety audits, or formal coordination between rivals remains to be seen. 
