Anthropic CEO warns of AI espionage threat from China

Anthropic CEO Dario Amodei has expressed concerns over potential espionage activities targeting top US artificial intelligence companies, warning that spies, likely from China, may be accessing valuable 'algorithmic secrets.'

Speaking at a Council on Foreign Relations event on Monday, Amodei highlighted the risk posed by China’s 'large-scale industrial espionage' and suggested that AI companies like Anthropic are almost certainly being targeted.

"Many of these algorithmic secrets, there are $100 million secrets that are a few lines of code," Amodei 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 increased support from the US government to mitigate this threat, though he did not specify the exact measures he believed were necessary.

In its recent recommendations to the White House’s Office of Science and Technology Policy (OSTP), the company advocated for the federal government to collaborate with AI industry leaders to strengthen security at frontier AI labs. This would include cooperation with US intelligence agencies and their allies.

Amodei’s comments align with his broader concerns about Chinese AI development. He has previously called for strict US export controls on AI chips to China and criticised Chinese AI lab DeepSeek, claiming it performed "the worst" on a bioweapons data safety test conducted by Anthropic.

His stance has drawn criticism from some members of the AI community, who argue that increased collaboration between the US and China on AI development could help prevent an AI arms race. Critics warn that such a race could lead to the creation of AI systems so powerful they become difficult for humans to control.

The debate reflects growing tensions around AI development, with security, ethics, and international cooperation at the forefront of discussions among industry leaders and policymakers.