At the end of August, the California federal appeals court sided with Anthropic in its dispute with the Pentagon over national security threats, but the conflict group still has to wait for the Washington court’s decision to reach a turning point. In fact, this did not happen, as the capital’s court upheld humanity’s existing status as a source of national security threats.
Image source: Human
The Defense Department first awarded the status to the startup in March, when the company clashed with Anthropic’s founders, who objected to the use of its artificial intelligence models to select targets on the battlefield and spy on U.S. citizens. Anthropic lost its contract with the U.S. military, and its contractors were banned from using the startup’s developments in official events commissioned by the Pentagon. Yesterday, the federal appeals court in Washington support Anthropic retained its supply chain threat status at two-thirds, rejecting the company’s argument that the Defense Department’s March decision was arbitrary, illegal and unconstitutional.
“The Department has provided compelling evidence that Crowder’s continued integration into the Department’s information systems, including the involvement of contractors, poses a legally recognized national security risk,” – stated in the decision of Federal Judge Gregory Casas. The decision was upheld by Judge Naomi Rao, appointed by President Trump, with only a third judge, Karen LeCraft Henderson, dissenting.
Anthropic founder Dario Amodei’s September call for a voluntary slowdown in artificial intelligence development angered U.S. President Donald Trump, who was not invited to a government dinner with Chinese leader Xi Jinping this week. On Monday, Trump said on his social network Truth Social: “The Trump administration is blocking artificial intelligence from people trying to do bad things or potentially doing bad things — like Anthropic’s Dario, who is now pretending to be a perfect little angel, and we will continue to do that.”.
Let’s recall that last month, a federal court in California declared Anthropic’s designation as a national security threat unlawful, but Anthropic appealed the Pentagon’s decision on two separate occasions, so the startup had to challenge the designation twice in order to be completely vindicated. This week, a U.S. court in Washington ruled against Anthropic in a second appeal, so the Pentagon appointment is valid and should remain relevant. Anthropic representatives told CNBC that they disagreed with the court’s decision, citing the California court’s decision in its favor, and therefore the company reserves the right to seek further opportunities to review the current status of its relationship with the Pentagon.
The Washington court gave Anthropic time to consider a second appeal before the same panel or a hearing before an expanded panel, so the current ruling has not yet taken effect. Additionally, Anthropic may ask the U.S. Supreme Court to hear the case. In his argument, Judge Casas echoed Defense Secretary Pete Hegseth’s concerns about the dangers of giving control of artificial intelligence models to a party that would not allow the secretary of defense and the president to make decisions related to military operations alone. As the judicial panel chair agreed with Hegseth’s argument, anthropic interference could lead to unnecessary manipulation or abrupt shutdown of AI models. Judge Casas said the Secretary of Defense’s desire to retain full authority in this area does not violate the Supply Chain Security Act or the U.S. Constitution.
If you find an error, select it with your mouse and press CTRL+ENTER.
