Trump Administration Retaliated Against Anthropic for Refusing to Remove Minimal Restrictions on Military’s Use of Its AI Tools

NEWS & RESEARCH

The Trump administration banned federal use of the AI company Anthropic, designating it a "supply chain risk" after the company refused to remove mild safety guardrails on military applications. While Anthropic has not opposed current U.S. military operations in Venezuela and Iran, it sought to prevent future use of its technology in lethal autonomous weapons and expanded mass surveillance. In response, President Trump and Defense Secretary Pete Hegseth condemned the restrictions as left-wing "virtue signaling" and demanded Anthropic permit "all lawful use" of its AI.

SOURCES: WIRED | The Guardian

ANALYSIS & OPINION

Former White House senior AI policy adviser Dean Ball criticized the supply chain risk designation—which bars any company doing business with the military from working with Anthropic—as essentially sanctioning an American company. The Guardian noted the administration's retaliation was an overreaction, given that Anthropic and the government generally align on acceptable use cases. Furthermore, a federal judge remarked that the Pentagon appeared to be "attempting to cripple" Anthropic over a contract dispute.

SOURCES: WIRED | The Guardian | CNBC

HOW TO FIX IT

Federal action:

  • Support calls for congressional action to reverse the decision to designate Anthropic a supply chain risk.

  • Support demands for answers from the DOD on its reasons for the supply chain designation, as well as answers from AI companies on the terms of their agreements with the DOD that could enable mass surveillance.

  • Reintroduce and pass the Fourth Amendment Is Not For Sale Act to prohibit law enforcement and intelligence agencies from purchasing sensitive protected information from third-party sellers.

  • Reintroduce and pass the AWARE Act, requiring the Pentagon to report to Congress on risk assessments of the use of AI including “privacy risks, and potential for civilian harm.”

  • Investigate “the DOD’s use of autonomous and semi-autonomous weapons systems.”

  • Adapt laws “governing the collection, analysis, and retention of private communications and other data” to “address the privacy risks of training, developing, and using AI in national security systems.”

Legislation: S.2576 - Fourth Amendment Is Not For Sale Act | S.5239 - AWARE Act

Previous
Previous

Trump Backed Ellisons’ Purchase of Paramount with Assurance of Sympathetic Coverage

Next
Next

Trump Directed DOJ to Investigate Fundraising Platform for Democrats