“We are grateful that the court has recognized the need to quickly resolve these issues,” said Anthropic. On the other hand, the U.S. government praised Wednesday’s decision as a “resounding victory,” according to Todd Blanche, the number 2 official at the Department of Justice. He added that “ethical decisions belong to the commander-in-chief and not a technology company,” emphasizing that “the military needs full access to Anthropic’s models if their technology is integrated into our sensitive systems.”
The decision upholds a measure aimed at forcing Pentagon subcontractors to certify that they do not use Anthropic’s technologies, including the chatbot Claude. This measure was announced on February 27 in response to Anthropic’s refusal to allow their AI tools to be used for mass surveillance of American citizens and to develop fully autonomous weapons. Previously, only non-U.S. companies had been targeted with this designation, leading to a cascade of public contract cancellations.
Context: Anthropic has been involved in a legal battle with the U.S. government regarding the use of its AI technologies by the military.
Fact Check: The U.S. government is seeking to restrict the use of Anthropic’s technology due to concerns over national security and ethical issues.
Regulatory Framework for Sanctions
The effectiveness of this sanction, which falls under the federal procurement code, is subject to debate, with some jurists arguing that the necessary regulatory texts for its implementation are still lacking. In a parallel legal action on March 26, Anthropic had previously won a victory against a similar sanction issued under military authority. In that case, a judge in San Francisco suspended a directive from Defense Secretary Pete Hegseth designating Anthropic as a “risk” for classified Pentagon operations.
The Californian judge, believing that these sanctions likely violated Anthropic’s freedom of expression, also suspended a directive from Donald Trump ordering all federal agencies to stop using their technologies. Following this, the access platform for federal agency AI, USAi.gov, restored access to Anthropic’s models on April 3.
The Pentagon informed Anthropic that they would continue to use their models, the only ones authorized for U.S. military classified information, until at least September 3, 2026. Despite being labeled a security risk, OpenAI had secured a contract with the Pentagon to replace Anthropic. The government has appealed the Californian court’s decision, but the hearing will not take place until at least June.
Context: Anthropic has faced legal challenges regarding restrictions on its technology’s use by U.S. government agencies.
Fact Check: Anthropic has been granted temporary relief from certain sanctions, allowing continued use of its AI models by the Pentagon for classified information.
“Exponential” Revenue Growth
In their decision on Wednesday, the Washington judges concluded that the government’s interests outweighed those of Anthropic. While Anthropic may suffer “likely irreparable harm,” it is primarily financial, whereas the government’s stake concerns “the security of Pentagon operations in an active military conflict.” Despite acknowledging the contracts lost by Anthropic, the court suggests that the company may have also benefited from the situation.
The judges reference statements from Dario Amodei, Anthropic’s CEO, to his employees, where he said “the general public and the media see us as heroes (we are number 2 in the App Store!),” and an article from Digiday suggesting that Anthropic’s opposition to the Pentagon “could turn out to be the best marketing investment in Silicon Valley in years.”
