Particle.news
Download on the App Store

Judge Blocks Pentagon’s Anthropic Blacklist, Citing Likely First Amendment Violations

The case could reset how the government polices AI vendors that set safety limits on military use.

Overview

  • Judge Rita Lin, who granted a preliminary injunction Thursday, temporarily blocked the Pentagon’s “supply‑chain risk” label and the White House’s government‑wide ban with a one‑week stay for an appeal.
  • The order restores conditions to February 27 and makes clear it does not force the Defense Department to use Anthropic while allowing agencies to move away from Claude only if they follow the law.
  • Lin wrote that the actions were likely unlawful, called them arbitrary and capricious, and said the record supports that officials punished Anthropic for speaking publicly about the contract dispute.
  • The fight grew from talks over a $200 million deal after Anthropic refused uses tied to mass domestic surveillance and fully autonomous lethal weapons, while government lawyers argued the company could not be trusted to keep models available as directed.
  • The “supply‑chain risk” tag is a tool usually aimed at foreign adversaries, and after negotiations collapsed the Pentagon turned to OpenAI as related litigation continues in California and the D.C. Circuit.