Anthropic CEO Dario Amodei Outlines Plan to Slow Frontier AI, With OpenAI Set to Follow

Anthropic chief executive Dario Amodei used a new blog post to argue that the industry should put the brakes on the race to build ever more capable AI systems, setting out three broad approaches for doing so rather than a single prescription.

Amodei said Anthropic is committing to one of those strategies on its own, without waiting for rivals to match the move. OpenAI CEO Sam Altman responded by indicating his company would take the same path.

Two developments convinced Amodei that a more cautious posture is warranted, according to the post: a hack involving OpenAI and HuggingFace, and what he described as drastically faster recent progress in AI, particularly models' growing ability to help build the next generation of the technology.

He wrote that the industry must slow the rate at which it improves the capabilities of AI models, adding that progress will still appear fast and that the time gained should be used wisely.

The post did not directly address the resignation of Anthropic researcher Jacob Coxon, who said this week he is leaving the company over concerns that leading AI labs are gambling with lives and that the people building the technology believe it could kill everyone by the end of the decade. Others at Anthropic have raised similar concerns, and the episode has sharpened an already intense debate over AI safety and alignment.

Amodei's call to slow the frontier follows earlier comments from Altman suggesting the industry may need to pace its development, a rare point of agreement between the two rival chief executives.