Anthropic’s CEO Dario Amodei is concerned about spies, likely from China, obtaining expensive “algorithmic secrets” from top AI companies in the U.S. He is urging the U.S. government to intervene.
Speaking at a Council on Foreign Relations event on Monday, Amodei mentioned China’s history of “large-scale industrial espionage” and the likelihood of AI companies like Anthropic being targeted.
“Many of these algorithmic secrets, there are $100 million secrets that are a few lines of code,” he said. “And, you know, I’m sure that there are folks trying to steal them, and they may be succeeding.”
Amodei stressed the importance of further assistance from the U.S. government to mitigate this risk, but didn’t specify the type of help needed.
Anthropic chose not to respond to TechCrunch’s request for comments on these remarks but referred to Anthropic’s recommendations to the White House’s Office of Science and Technology Policy (OSTP) from earlier this month.
In the submission, Anthropic proposes that the federal government collaborates with leading AI industry players to enhance security at advanced AI labs, including engaging with U.S. intelligence agencies and their partners.
Amodei’s views align with his critical stance on Chinese AI development. He has advocated for stringent U.S. export controls on AI chips to China and highlighted DeepSeek’s poor performance on a critical bioweapons data safety test conducted by Anthropic.
Amodei’s concerns, as outlined in his essay “Machines of Loving Grace” and other platforms, revolve around China’s utilization of AI for authoritarian and military purposes.
This stance has drawn criticism from some in the AI community who argue for increased collaboration between the U.S. and China in AI to prevent an arms race leading to the development of a system beyond human control.
