Dario Amodei, CEO of AI research company Anthropic, has raised concerns that foreign intelligence agents, particularly from China, may be stealing critical algorithmic data from leading U.S. artificial intelligence firms.
Speaking at a Council on Foreign Relations event, Amodei warned that AI companies like Anthropic are likely being targeted for their proprietary technology, which he described as “algorithmic secrets” worth hundreds of millions of dollars.
According to Amodei, China has a well-documented history of large-scale industrial espionage, and AI is a prime target due to its strategic importance in technological and military development. He stated that some AI breakthroughs consist of “a few lines of code” that hold immense value, making them attractive to foreign intelligence services. “I’m sure that there are folks trying to steal them, and they may be succeeding,” he said.
Amodei called for increased U.S. government intervention to protect AI companies from such threats. However, he did not specify what measures he believed were necessary to counter espionage.
Anthropic declined to comment further on Amodei’s remarks but pointed to a submission it made earlier this month to the White House’s Office of Science and Technology Policy (OSTP). In this submission, the company urged the U.S. government to work more closely with AI industry leaders to enhance security at AI research facilities. This could include partnerships with intelligence agencies and allied governments to safeguard sensitive research.
Amodei has previously taken a strong position against China’s AI development. He has supported U.S. export restrictions on AI chips destined for China and has criticised Chinese AI models for their potential misuse. He singled out the Chinese AI company DeepSeek, saying that it performed poorly in a bioweapons safety assessment conducted by Anthropic.
His broader concerns, outlined in his essay Machines of Loving Grace, focus on the possibility that China could use AI for authoritarian control and military applications. These warnings align with growing U.S. fears over the role of AI in national security.
However, not all AI experts agree with Amodei’s position. Some argue that increased collaboration between the U.S. and China on AI development is necessary to prevent an uncontrolled technological arms race. They fear that without cooperation, both countries could develop AI systems so powerful that they become difficult to manage or regulate, leading to risks for global stability.
The debate over AI security and international cooperation continues, with growing pressure on the U.S. government to define its policy on AI-related espionage and national security risks.
Read also:
France’s AI Chatbot Lucie Pulled Offline After Embarrassing Errors
