Anthropic and the Department of Defense: A Deepening Dispute Over AI Access and National Security

A legal battle is escalating between Anthropic, a leading artificial intelligence developer, and the U.S. Department of Defense (DoD) over access to its AI technology. The DoD has designated Anthropic as a supply chain risk, a move Anthropic is challenging in court, alleging the government is retaliating against the company’s stance on responsible AI deployment. The core of the dispute centers on the DoD’s desire for unrestricted military use of Anthropic’s Claude AI model, and Anthropic’s concerns about potential misuse, particularly regarding autonomous weapons and mass surveillance.

The Roots of the Conflict

The conflict began in late February 2026, when Defense Secretary Pete Hegseth and President Trump announced the DoD was cutting ties with Anthropic after the company refused to grant unrestricted military access to its AI technology [WIRED]. Anthropic, which has been proactively deploying its models within the U.S. Government and intelligence community since 2025 [Anthropic – Statement from Dario Amodei], argues the DoD’s actions are a violation of its First Amendment rights.

Anthropic’s Declarations and Counterclaims

In sworn declarations submitted to a California federal court on March 22, 2026, Anthropic officials refuted key claims made by the government. Sarah Heck, Anthropic’s Head of Policy, stated that the company never sought an approval role over military operations, nor did the DoD raise concerns during negotiations about Anthropic potentially disabling or altering its technology mid-operation [WIRED]. Heck too pointed to an email from the DoD’s Under Secretary Emil Michael, dated March 4, 2026, suggesting the two sides were “particularly close” to alignment on issues of autonomous weapons and mass surveillance, contradicting subsequent public statements from the DoD [WIRED].

Thiyagu Ramasamy, Anthropic’s Head of Public Sector, addressed concerns about potential interference with military operations. He asserted that once Claude is deployed within a government-secured system, Anthropic has no access or control over it, and any changes would require explicit Pentagon approval [WIRED]. Ramasamy also highlighted that Anthropic employees undergo rigorous U.S. Government security clearance vetting, a standard not commonly met by other AI companies operating in classified environments.

The Government’s Position

The Department of Justice, representing the DoD, countered that Anthropic’s refusal to allow all lawful military uses of its technology was a business decision, not protected speech. They maintain the supply-chain risk designation is a legitimate national security measure, not retaliation for the company’s views [WIRED]. The government argues that Anthropic’s actions could potentially jeopardize national security and that the designation does not violate the company’s First Amendment rights.

Broader Implications and Future Outlook

This dispute highlights the growing tension between the rapid advancement of AI technology and the need for responsible deployment, particularly within the defense sector. Anthropic’s proactive stance on AI safety, including cutting off access to firms linked to the Chinese Communist Party and advocating for export controls [Anthropic – Statement from Dario Amodei], underscores the complex geopolitical considerations surrounding AI development. A hearing is scheduled for March 24, 2026, before Judge Rita Lin in San Francisco to determine whether to grant Anthropic a reprieve even as the litigation is resolved [WIRED]. The outcome of this case could set a significant precedent for how the government regulates access to advanced AI technologies and the balance between national security and the principles of responsible AI development.

Anthropic has a strong working relationship with the Department of War, supporting applications such as intelligence analysis, modeling and simulation, operational planning, and cyber operations [Anthropic – Where things stand with the Department of War].