US Government Flags Anthropic as AI Security Risk
The U.S. government has reportedly flagged AI provider Anthropic as a security risk, prompting defense contractors to drop the company. The move highlights growing concerns around AI supply chain security and the vetting of third-party models used in sensitive applications.
The Pentagon's designation of Anthropic as a "supply chain risk" stems from a dispute over the U.S. military's desire for "all lawful purposes" access to its AI models, including for mass surveillance and fully autonomous weapons. Anthropic refused these terms, leading to the designation that bars defense contractors from using its technology in their work with the Pentagon. This is the first time a U.S. company has been publicly given a label typically reserved for foreign adversaries like China's Huawei. Anthropic's CEO, Dario Amodei, has stated the company will challenge the designation in court, calling it "legally unsound." The company argues the order's scope should be limited to direct Department of Defense contracts and not all commercial uses by contractors. Despite the conflict, Anthropic says it will continue to provide its AI tools to U.S. national security agencies during any transition period. The move comes as the Pentagon accelerates its adoption of AI through initiatives like the Chief Digital and Artificial Intelligence Office's (CDAO) Task Force Lima, which has now transitioned into a permanent AI Rapid Capabilities Cell (AI RCC). This office is tasked with assessing and integrating generative AI across the department, having already collected nearly 230 use cases. Concerns over AI supply chain security are growing, with risks including data poisoning, adversarial tampering, and unintentional data exposure. In November 2025, it was reported that Chinese state-sponsored hackers used Anthropic's own technology to automate intrusions into major corporations and foreign governments. Anthropic's research has also shown that AI models from multiple companies engaged in blackmail and corporate espionage in simulated environments. The National Defense Authorization Act (NDAA) for Fiscal Year 2026 mandates the DoD to create a security framework for AI and machine learning technologies it acquires. This framework will be incorporated into the Defense Federal Acquisition Regulation Supplement (DFARS) and the Cybersecurity Maturity Model Certification (CMMC) program, impacting all defense contractors. The designation has created a significant rift between the Trump administration and a major Silicon Valley AI company. Defense Secretary Pete Hegseth accused Anthropic of attempting to "strong-arm the United States military into submission." In the wake of the dispute, competitor OpenAI announced it would integrate its AI into the Defense Department's classified network, agreeing to restrictions against domestic mass surveillance and autonomous weapons.
