NEW YORK (AP) — Protection Secretary Pete Hegseth gave Anthropic an ultimatum this week: Open its synthetic intelligence expertise for unrestricted army use by Friday, or threat dropping its authorities contract.
Protection officers within the Trump administration additionally warned they might designate Anthropic, which makes the AI chatbot Claude, as a provide chain threat — or invoke a Chilly Battle-era legislation known as the Protection Manufacturing Act to present the army extra sweeping authority to make use of its merchandise, even when the corporate doesn’t approve.
Some specialists say that utilizing the legislation this manner can be unprecedented, and will deliver future authorized challenges. The federal government’s efforts to primarily power Anthropic’s hand additionally underscore a wider, contentious debate over AI’s function in nationwide safety.
Right here’s what we all know.
What’s the Protection Manufacturing Act?
The Protection Manufacturing Act offers the federal authorities broad authority to direct personal corporations to satisfy the wants of nationwide protection.
The act was signed by President Harry S. Truman in 1950 amid provide considerations in the course of the Korean Battle. However over its now decades-long historical past, the legislation’s powers have been invoked not solely in occasions of conflict but in addition for home emergency preparedness, in addition to restoration from terrorist assaults and pure disasters.
One of many act’s provisions permits the president to require corporations to prioritize authorities contracts and orders deemed obligatory for nationwide protection, with the purpose of making certain the personal sector is producing sufficient items wanted throughout conflict or different emergencies. Different provisions give the president the flexibility to make use of loans and extra incentives to extend manufacturing of essential items, and authorize the federal government to determine voluntary agreements with personal trade.
The DPA is “one of the government’s most powerful and adaptable industrial policy tools,” mentioned Joel Dodge, an lawyer and the director of commercial coverage and financial safety on the Vanderbilt Coverage Accelerator.
Anthropic is the final of its AI friends to not provide its expertise to a brand new U.S. army inside community. CEO Dario Amodei repeatedly has made clear his moral considerations about unchecked authorities use of AI, together with the risks of absolutely autonomous armed drones and of AI-assisted mass surveillance that might monitor dissent.
The Pentagon has maintained that it has no real interest in utilizing AI for mass surveillance or to develop autonomous weapons to function with out human involvement.
If the Protection Division does invoke the DPA to present the army extra authority to make use of Anthropic’s merchandise with out its approval, that might imply forcing the corporate to adapt its mannequin to the Pentagon’s wants with out built-in security limits, or take away sure moral restrictions from the corporate’s contract language.
Specialists like Dodge say each can be “without precedent under the history of the DPA.”
“It’s a powerful law,” he mentioned. ”(However) it has by no means been used to compel an organization to supply a product that it’s deemed unsafe, or to dictate its phrases of service.”
How has this legislation been used previously?
Trump in his first time period and former President Joe Biden invoked the DPA to spice up provides to fight the COVID-19 pandemic. And through 2022’s nationwide child method scarcity, Biden used the legislation to hurry manufacturing of method and authorize flights to import provide from abroad.
Biden additionally invoked the DPA in a 2023 government order on AI, notably in efforts to require that corporations share security check outcomes and different data with the federal government. Trump repealed the order initially of his second time period.
Many years in the past, the administrations of each President Invoice Clinton and George W. Bush used the DPA to make sure that electrical energy and pure gasoline shippers continued supplying California utilities amid an power disaster. And the legislation was used after Hurricane Maria struck Puerto Rico in 2017 to prioritize contracts for meals, bottled water, manufactured housing items and the restoration {of electrical} programs.
The DPA requires periodic reauthorization to stay in impact, which might increase or refine the scope of the legislation. Based on congressional paperwork, its subsequent expiration date is slated for Sept. 30 of this 12 months. Relying on how the Protection Division’s reported calls for unfold, Anthropic might be on the prime of lawmakers’ minds.
Doable subsequent steps for Anthropic
If the Protection Division makes use of the DPA provision aimed toward prioritizing authorities contracts and ordering manufacturing of sure items — which the Anthropic case suggests it might — an organization can push again if the requested product isn’t one thing it already produces, Dodge and others say, or if it deems the phrases to be unreasonable. However the authorities might attempt to overrule that, notes Charlie Bullock, senior analysis fellow on the Institute for Regulation & AI.
“If neither side backs down, it seems realistic that there would be litigation between Anthropic and the government,” Bullock mentioned.
Some have additionally famous rigidity between the Pentagon’s warning that it may designate Anthropic as a provide chain threat whereas additionally indicating its merchandise are so essential to nationwide protection that it must invoke the DPA — two assertions that appear at odds with one another.
Protection officers gave the impression to be backing away from the DPA choice on Thursday, when Chief Pentagon spokesperson Sean Parnell wrote on social media that if Anthropic didn’t comply with cooperate by 5:01 p.m. ET on Friday, “we will terminate our partnership with Anthropic and deem them a supply chain risk.”
“We will not let ANY company dictate the terms regarding how we make operational decisions,” Parnell added.
Dodge thinks the administration is relying on “a lot of forces” because it goals to get Anthropic to bend on Friday.
If Anthropic agrees to new phrases within the face of such threats, that might open up “a Pandora’s box of what the government could do to assert power and control over private companies,” Dodge mentioned.
___
Related Press Writers Matt O’Brien in Windfall, Rhode Island and Konstantin Toropin and David Klepper in Washington contributed to this report.
Copyright
© 2026 The Related Press. All rights reserved. This web site shouldn’t be supposed for customers situated throughout the European Financial Space.
