Victory for Anthropic: the US court blocked its inclusion in the blacklist

In the lawsuit it submitted to the court, Anthropic claimed that US Defense Secretary Pete Hegseth exceeded his authority when he defined the company as a “risk to the supply chain” of national security. This is a classification that the government can apply to companies that expose military systems to possible penetration or sabotage by hostile elements.

Hesseth’s move, which was unprecedented, came after Anthropic refused to lift two restrictions it had placed on military use of the Claude models: use for mass surveillance of American citizens and use of fully autonomous weapons. Following the classification, Anthropic was blocked from certain military districts, and its executives claimed that the move could cost the company billions of dollars in lost business and damage its reputation.

The ruling was given by federal judge Rita P. Lynn, who was appointed to the position by former President Joe Biden. In a 59-page decision, the court stated that the Pentagon’s decision was “illegal and baseless”, and that reliance on national security considerations does not give the government a free hand to punish its critics.

The court ruled that the measures against Anthropic violated her right to freedom of expression enshrined in the US Constitution. It was further determined that Anthropic was not given an opportunity to challenge the classification before it was imposed on her, contrary to her right to a fair process. In addition, the court stated that the decision was arbitrary.

This is the first time that an American company has been publicly identified as a risk to the supply chain by virtue of a government procurement law designed to protect military systems from foreign sabotage.

Months of confrontations with the Trump administration

The ruling comes after months of conflict between Anthropic and the Trump administration. As we previously published in Globes, in June the US Department of Commerce ordered the company to limit access to two of its advanced models, Claude Fable 5 and Claude Mythos 5, for foreign citizens, due to national security concerns. As a result, Anthropic disabled access to the models, and the administration’s move drew harsh reactions in the industry.

However, less than three weeks later, the administration withdrew the restrictions. In early July, the Department of Commerce lifted the export restrictions and Anthropic began to put the models back into operation. In return, the company agreed to work with the government on protocols and standards in the field of security and to act proactively to identify security risks, as reported in Globes.

The legal battle is not over

At the heart of the current conflict is the question of the limits of the military use of artificial intelligence. Anthropic claims that AI models are not yet reliable enough to be used safely in autonomous weapons systems, and also opposes their use for domestic surveillance in the US on civil rights grounds.

The Pentagon argued on the other hand that private companies should not be allowed to limit the way in which the military uses the technology it has acquired. The US Department of Justice argued in court that Anthropic’s refusal to remove the restrictions on the use of Claude could create uncertainty in the Pentagon about how the model can be used and even endanger military systems during operational activity.

According to the government, the decision against the company stemmed from its refusal to accept the terms of the engagement and not from its positions regarding the safety of artificial intelligence.

Anthropic welcomed the ruling and stated that it continues to focus on beneficial cooperation with the government in order to harness artificial intelligence for the needs of national security.

However, the legal battle is not over yet. Anthropic is pursuing another lawsuit in Washington against the Pentagon’s separate classification, which could also exclude civilian government districts.