Pentagon: Anthropic's foreign workforce is a China risk
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Pentagon undersecretary Emil Michael stated in a March 17 court declaration that Anthropic employs many foreign nationals including from China, raising the risk those employees could be compelled under the PRC's National Intelligence Law.
- The filing singles out Anthropic from other major U.S. AI labs, arguing their foreign-workforce risks are reduced by 'the technical and security assurances of the other labs' leadership' and consistent responsible behavior with the Defense Department.
- The Pentagon acknowledged it still relies on Anthropic and is willing to extend the deadline for offboarding the company's tools, even as it argues broader national security concerns beyond surveillance and autonomous weapons.
- Anthropic was an early adopter of operational security measures like research compartmentalization and audit trails, was the first AI lab to partner with the Pentagon, and last year disrupted what it called a first-of-its-kind Chinese AI-orchestrated cyber espionage campaign on its platform before banning the PRC from its services.
- Samuel Hammond of the Foundation for American Innovation called insider threats 'genuine and tricky' but said Anthropic is 'widely considered to be the most serious and proactive about policing insider threats from foreign nationals.'
- Chinese-origin researchers make up roughly 38-40% of top AI talent at U.S. institutions as of 2023, according to a global AI talent tracker cited by Axios, underscoring how broadly the Pentagon's concern could apply across the industry.
- A court hearing on whether to grant Anthropic temporary relief from the supply-chain-risk designation, which the company is asking courts to overturn, is set for March 24.
Why it matters: If a court accepts the Pentagon's national security framing, it could lock in Anthropic's supply-chain-risk designation and accelerate offboarding from defense work, even though the filing concedes the department still depends on Anthropic's tools. The reasoning also extends well beyond Anthropic: with Chinese-origin researchers making up an estimated 38-40% of top U.S. AI talent, the same workforce-vulnerability claim could be aimed at virtually any major lab.
Ask SkimNews
