Anthropic and Pentagon Face Off in Court as Judges Appear Divided Over AI Blacklisting

Anthropic and Pentagon Square Off in High-Stakes Legal Battle

A federal appeals court in Washington, D.C., heard intense arguments on Tuesday as Anthropic and the Pentagon faced off over the U.S. Department of Defense's controversial blacklisting of the AI company 1

1

. The legal dispute erupted after Defense Secretary Pete Hegseth designated Anthropic a supply chain risk in March, a label historically reserved for foreign adversaries that effectively bars defense contractors from using the company's Claude AI models 1

1

. The designation came after months of failed negotiations between Anthropic and the DOD, with the Pentagon demanding unfettered access to Anthropic's technology for all lawful use while the AI company insisted on restrictions against autonomous weapons and mass domestic surveillance 1

1

.

Judges Karen Henderson, Gregory Katsas, and Neomi Rao presided over the Pentagon AI suit, with each side receiving 15 minutes to present their case 1

1

. Anthropic attorney Kelly Dunbar faced nearly an hour of intense questioning, arguing that Hegseth "turned a powerful national security authority against an American company, and he did so to gain leverage in a contract dispute" 4

4

.

Appeals Court Judges Display Sharp Division Over Blacklisting

The three-judge panel revealed starkly different perspectives on the Anthropic blacklisting case. Judge Henderson, appointed by President George H.W. Bush, expressed strong skepticism about the Pentagon's rationale, stating: "For the life of me, I do not see any evidence of maliciousness" and called it a "spectacular overreach by the department" 2

2

. Henderson questioned whether the Pentagon had considered less intrusive means before resorting to the supply chain risk designation 5

5

.

However, Judges Katsas and Rao, both appointed by the Trump administration, appeared more sympathetic to national security concerns. Rao questioned what basis the court could have for second-guessing Hegseth's judgment, noting the government's reliance on the opaque nature of AI models 5

5

. Katsas highlighted the difficulty of Anthropic's usage policies for rapidly evolving technology, pointing out that that "AI three months from now will be totally different from the AI of today" 2

2

.

AI Safety and Ideological Retaliation at Center of Dispute

The AI company legal battle centers on fundamental questions about AI safety and government overreach. Anthropic refused to agree to the Pentagon's all lawful use standard, maintaining red lines around the use of its tools for mass domestic surveillance or the development of weapons that fire without human involvement 2

2

. The company argues it has no way to control its AI models once they're deployed in classified settings, making such assurances impossible to enforce 2

2

.

Anthropic contends the designation amounts to ideological retaliation for raising ethical concerns, with CEO Dario Amodei stating the company had "no choice" but to challenge the blacklisting in court 1

1

3

3

. The Pentagon claims it's unworkable to rely on Anthropic because the company might pull the plug at any time due to its "ideological" views around AI safety 2

2

.

National Security Concerns and Trust Issues Take Center Stage

Government lawyer Sharon Swingle argued that trust lies at the heart of the issue, pointing to a "very real prospect of new red lines" and an "increasingly hostile posture" in negotiations 2

2

. Swingle told the court that Anthropic clearly has the ability to interfere with the Pentagon's usage of Claude "for critical military operations," adding that "the failure of the model in active military operations could have catastrophic national-security consequences and put service members' lives at risk" 5

5

.

Dunbar countered that if the government doesn't trust Anthropic's model, it could simply decide not to do business with the company rather than applying a permanent legal disbarment that risks broader government-wide and commercial exclusions while damaging the company's reputation 2

2

. The blacklisting jeopardizes hundreds of millions in government contracts for the San Francisco-based company 3

3

.

What Happens Next for Anthropic and Defense Contractors

The judges did not indicate when they would issue a ruling on Anthropic's appeal 5

5

. A split decision between the federal appeals court in DC and a San Francisco court means Anthropic can't enter into new defense contracts but can continue its contracts with non-Pentagon agencies while litigation plays out 2

2

. Earlier this month, a federal judge in San Francisco ruled in Anthropic's favor and blocked the Pentagon from labeling the company as a supply chain risk 5

5

.

The case carries major implications for the government's working relationship with technology companies and federal procurement processes 4

4

. Notably, the DOD continued to use Anthropic's AI models to support its military operations in Iran even after the blacklisting, and President Donald Trump told CNBC last month that a deal between the DOD and the startup is "possible" 1

1

. This creates an awkward situation where the Pentagon treats a U.S. company as a national security threat while simultaneously relying on its technology to combat foreign adversaries 2

2

.