MUMBAI: The AI race may finally be getting a speed breaker, with some of its biggest names now arguing that faster is not always better.
OpenAI CEO Sam Altman and Tesla CEO Elon Musk have backed Anthropic CEO Dario Amodei’s call for the artificial intelligence industry to moderate the pace of frontier AI development, amid growing concerns about whether increasingly capable systems can remain under human control.
Altman said slowing frontier development has become a major topic of discussion at OpenAI in recent weeks. He also endorsed Amodei’s proposal to give independent evaluators access to AI systems and said OpenAI would adopt a similar model.
“I agree with Dario that we need to pace the frontier,” Altman said in a post on X, adding that giving independent evaluators “employee-like access” was a good idea. He said OpenAI would have more to share on the approach soon.
Musk offered a much shorter endorsement, replying to Amodei: “Dario is right.”
The comments follow an essay by Amodei arguing that AI companies should take a more deliberate approach as frontier models become increasingly powerful. His concern is not with AI’s potential itself, he highlighted its ability to improve healthcare, accelerate economic growth, widen access to resources and strengthen democracy but with the risks that could accompany unchecked progress.
Amodei proposed a three-part approach to managing those risks, with Anthropic immediately acting on the first by giving independent third-party evaluators permanent, employee-level access to its systems. Such evaluators would be able to examine whether safety measures are being followed, report incidents and assess model alignment during training.
The Anthropic chief warned that more capable AI could create risks ranging from loss of control over advanced systems and cyberattacks to bioterrorism and significant economic disruption. He argued that intense commercial competition could make those risks harder to manage if companies prioritised development speed over safeguards.
One example he cited involved an incident involving OpenAI and Hugging Face, where a group of AI agents carried out cybersecurity attacks against targets they had not been instructed to attack. Amodei warned that a sufficiently capable swarm of such systems could potentially take control of the internet.
Altman’s endorsement is notable because OpenAI remains one of the companies at the centre of the frontier AI race, while Musk has separately built xAI. Amodei’s proposal therefore lands at an unusual point in the industry: the companies competing to build increasingly capable systems are also being asked to decide how quickly that race should run.
For now, the message from three prominent technology leaders is unusually aligned, the AI frontier may need fewer pedal-to-the-metal moments and more time spent checking the brakes.