Pentagon Can Blacklist Anthropic’s AI Technology
New York — The Department of Defense has the authority to blacklist Anthropic’s AI technology, according to a recent ruling from a federal appeals court. This decision poses a significant setback for Anthropic, particularly after a previous California court ruling in August that partially invalidated the Trump administration’s designation of the company.
In early 2023, the Trump administration labeled Anthropic as a “supply chain risk,” which prevented federal agencies and contractors linked to the military from engaging in business with the company. This designation was prompted by Anthropic’s refusal to let the Pentagon utilize its AI technology for all lawful purposes, particularly due to the company’s position against involvement in autonomous weapons and the surveillance of U.S. citizens.
The recent ruling from the D.C. Circuit Court confirms that Anthropic’s AI, known as Claude, will still be barred from use within the Pentagon. However, the earlier California decision permits other government agencies and contractors to collaborate with Anthropic, as the government has opted not to contest this ruling.
This Pentagon designation considerably jeopardizes a major portion of Anthropic’s enterprise business. Consequently, contractors, suppliers, or partners working with the military are prohibited from doing business with Anthropic, including access to its premium version of Claude. The company is banking on its strength in enterprise AI to promote a potential initial public offering (IPO) worth trillions, which might occur later this year.
In response to the designation, Anthropic filed a lawsuit, claiming the government violated the First Amendment. The company argued that the supply chain risk (usually applied to firms connected to foreign adversaries) was legally inappropriate for an American company.
Nevertheless, the federal Court of Appeals for the District of Columbia dismissed Anthropic’s claims, asserting that the Pentagon had sufficient grounds to determine that incorporating Claude into its systems could evoke national security concerns. One judge on the panel expressed dissent, arguing there was no substantial justification for fearing that Anthropic could misuse its technology for security threats.
An Anthropic spokesperson expressed disagreement with the ruling, stating, “Another federal court has already held the government’s parallel designation unlawful. We remain confident in our position and are considering all options, including further review.”
Anthropic may consider asking the entire D.C. appeals court for an “en banc” rehearing, ensuring a ruling from all 11 judges, or it could take its case directly to the Supreme Court.
It’s noteworthy that the Pentagon’s various designations led to the contrasting rulings from the D.C. and California courts, which were informed by different statutory authorities. The D.C. court did not overturn the California ruling.
In light of the ruling, the Pentagon expressed satisfaction, with Undersecretary Emil Michael declaring that “the hammer of justice has smashed [Anthropic’s] arguments… Warfighters will sleep better knowing that no private company will insert their opinions in the chain of command.”
