“Confirmed Judges, Confirmed Fears” is a blog series documenting the harmful impact of President Trump’s judges on Americans’ rights and liberties. It includes judges nominated in both his first and second terms.
What’s at stake in this case?
Anthropic tried to invalidate the blacklisting of Anthropic and its AI tool Claude from use by the Department or its contractors because of restrictions Anthropic placed on the tool.
What happened in this case?
For several years, the Defense Department had used Anthropic’s AI product Claude. It wanted Anthropic to remove several restrictions the company had placed on Claude, including banning its use for “lethal autonomous warfare” without human involvement or for “mass surveillance” of Americans. When the company refused to remove these restrictions, Defense blacklisted Anthropic starting in March 2026, banning the use of Claude by Defense personnel or by private contractors doing work for Defense. After exhaustion of internal remedies, Anthropic filed suit in the DC Circuit, claiming that the blacklisting was not authorized by law and was unconstitutional.
In September, Trump judges Greg Katsas and Neomi Rao issued a 2-1 decision that upheld the blacklisting of Anthropic in Anthropic v Department of War [sic]. George HW Bush judge Karen LeCraft Hendeson dissented. Katsas’ opinion maintained that the Department had “ample support” for its conclusion that using Claude would pose a “national security risk” that justified the blacklisting under the law.
Why did Judge Henderson dissent?
Judge Henderson strongly disagreed with the majority’s rejection of the “narrower reading” of the procurement law at issue, under which blacklisting is appropriate only if Anthropic had committed “intentionally subversive acts” executed through “deceptive means.” She concluded that the law does not allow blacklisting based on a contractor’s “honest and upfront enforcement of restrictions” like those Anthropic promulgated that are simply “disfavored by the government.”
Why is this decision harmful?
The Trump judge ruling by Katsas and Rao in this case obviously harms Anthropic, and also makes it easier for the Trump administration to insist on requiring AI contractors to facilitate government use of systems that remove human judgment from war-related decisions and to conduct mass surveillance. It also sets a bad precedent concerning interpretation of government procurement contracts. In addition, the decision illustrates the importance of our federal courts to health, welfare and justice and the significance of having fair-minded judges on the federal bench. The case may well be taken to the full DC Circuit.