Particle.news
Download on the App Store

Pentagon Threatens to Blacklist Anthropic as Deadline to Drop AI Use Limits Arrives

The dispute turns on Anthropic’s refusal to allow mass surveillance or autonomous weapons despite a push for unrestricted military use.

Overview

  • U.S. Defense Secretary Pete Hegseth issued an ultimatum expiring Friday evening for Anthropic to relax its user terms or face blacklisting.
  • The Pentagon seeks the freedom to decide how it employs Anthropic’s models, while the company maintains bans on mass surveillance and autonomous weapon use.
  • Officials have warned they could also restrict other defense contractors from working with Anthropic as additional leverage.
  • Anthropic previously secured Pentagon work worth up to $200 million and its Claude model was approved for the most secret military networks.
  • Reporting linked Claude’s use via Palantir to a January operation in Caracas targeting Nicolás Maduro, a disclosure that heightened tensions; Anthropic denies objecting to that use.