The U.S. Department of Defense fired back at Anthropic on Tuesday, calling the AI company an “unacceptable risk to national security” in its first formal court response to Anthropic’s lawsuits. TechCrunch AI reports the DOD filed a 40-page brief in a California federal court, escalating what has become the most significant clash between the AI industry and the U.S. military.
The core of the Pentagon’s argument: Anthropic might disable or alter its AI models during active military operations if the company decides its internal safety boundaries (its “red lines”) are being crossed. That’s a striking claim, and it cuts to the heart of a question the AI industry has been circling for years: who gets the final say over how powerful AI systems are used in warfare?
How We Got Here
Last summer, Anthropic signed a $200 million contract with the Pentagon to deploy its technology inside classified systems. During contract negotiations, the company drew two lines. It didn’t want its AI used for mass surveillance of Americans, and it said the technology wasn’t ready for lethal weapons targeting or firing decisions.
The Pentagon’s position: a private company shouldn’t dictate how the military uses its tools. Defense Secretary Pete Hegseth responded last month by labeling Anthropic a supply-chain risk, a serious designation that triggered Anthropic’s lawsuits.
The Legal Battle Heats Up
Anthropic has asked the court to temporarily block the DOD from enforcing the supply-chain label. In its lawsuits, the company argues the Pentagon is:
- Infringing on its First Amendment rights
- Punishing it on ideological grounds for refusing an “all lawful use” contract provision
- Retaliating against protected expression
Constitutional rights lawyer Chris Mattei, a former Justice Department attorney, told TechCrunch AI that the DOD’s argument doesn’t hold up. He pointed out there’s been no investigation to support the claim that Anthropic might disable its models during warfighting.
“The government is relying completely on conjectural, speculative imaginings to justify a very, very serious legal step they’ve taken against Anthropic,” Mattei said.
He also noted the DOD failed to explain why Anthropic’s negotiating stance made it an “adversary” rather than simply a vendor the Pentagon didn’t want to work with. The department could have just ended the contract.
Industry Rallies Behind Anthropic
What’s remarkable is the breadth of support Anthropic has received. Companies that are direct competitors, OpenAI, Google, and Microsoft, along with their employees and legal rights groups have filed amicus briefs backing Anthropic. That’s nearly unheard of in an industry this competitive.
The reason is obvious: every AI company negotiates usage terms with customers. If the government can label a company a national security threat for setting boundaries in a contract, that precedent affects the entire sector.
Why This Matters
This case will likely define the relationship between AI companies and the U.S. military for years to come. A few things to watch:
- The precedent question. Can the government weaponize supply-chain designations against companies that negotiate contract terms it doesn’t like?
- The safety implications. Anthropic has been one of the loudest voices on AI safety. If safety guardrails become a legal liability in government contracts, other companies may quietly drop them.
- The First Amendment angle. If the court sides with Anthropic, it could establish that refusing certain contract terms constitutes protected expression.
CEO Dario Amodei has maintained that “Anthropic understands that the Department of War, not private companies, makes military decisions,” adding the company has never tried to limit use of its technology in an ad hoc manner.
A hearing on Anthropic’s request for a preliminary injunction is scheduled for next Tuesday. An Anthropic spokesperson said seeking judicial review is a “necessary step” to protect its business, customers, and partners.
This story is developing fast. For full details, including the complete court filings and additional legal analysis, check the original reporting at TechCrunch AI.