A federal appeals court in Washington, D.C., on Wednesday rejected a request by artificial intelligence startup Anthropic to temporarily block the U.S. Department of Defense from blacklisting it as it sues over the sanctions.
Previously, a San Francisco federal court judge approved a preliminary injunction filed by Anthropic in a related case at the end of last month, prohibiting the Trump administration from enforcing a ban on the use of the company's Claude artificial intelligence technology.
"We believe that a fair balance in this case favors the government," the appeals court said in its decision. "On the one hand, the risk of economic harm to a private company is relatively small; on the other hand, how and through whom the U.S. Department of Defense acquires vital artificial intelligence technology during a military conflict requires judicial oversight. Therefore, we deny Anthropic's motion to stay pending review of the merits of the case."
The U.S. Department of Defense declared Anthropic a supply chain risk in early March, meaning use of the company's technology could threaten U.S. national security. The designation requires defense contractors to demonstrate that they do not use Anthropic's Claude artificial intelligence model in their work with the military.
Anthropic then appealed to the appeals court, asking for a review of the Pentagon's decision, arguing that it was an act of retaliation, unconstitutional, arbitrary, capricious and inconsistent with procedures established by law.
In Wednesday's ruling, the court acknowledged that "Anthropic is likely to suffer a degree of irreparable harm" if the proceedings are not stayed, but that the company's interests "appear to be primarily financial." The ruling states that while the company claimed that the U.S. Department of Defense obstructed its right to free speech, "Anthropic has not demonstrated that its speech was suppressed during the proceedings."
Given the potential harm to Anthropic, the appeals court said "expedited proceedings are necessary."