Dario Amodei, the CEO of Anthropic, expressed his concerns that spies, particularly from China, are attempting to acquire valuable "algorithmic secrets" from leading AI companies in the U. S. , and he is urging the government to intervene. During his remarks at a Council on Foreign Relations event this past Monday, Amodei highlighted China’s reputation for extensive industrial espionage, indicating that companies like Anthropic are likely targets. “Many of these algorithmic secrets represent substantial financial value, with some being worth $100 million and consisting of just a few lines of code, ” he explained. “There are certainly individuals attempting to steal these secrets, and they may very well be succeeding. ” Amodei emphasized that increased assistance from the U. S. government to mitigate this threat is “very important, ” although he did not clarify the specific type of support needed. While Anthropic did not comment directly to TechCrunch regarding Amodei's statements, the company referred to its recommendations submitted to the White House’s Office of Science and Technology Policy (OSTP) earlier this month. In that submission, Anthropic suggested that the federal government collaborate with leaders in the AI industry to enhance security protocols at advanced AI labs, which could involve coordination with U. S.

intelligence agencies and their global partners. Amodei's remarks align with his critical view of Chinese advancements in AI. He has advocated for stringent U. S. export controls on AI chips to China and noted that DeepSeek performed poorly on a significant bioweapons data safety assessment conducted by Anthropic. His concerns, articulated in his essay “Machines of Loving Grace” and other forums, focus on the potential for China to exploit AI for authoritarian and military ends.