Pentagon stops using Anthropic AI tools following national security risk designation

The United States Department of Defence has ceased using artificial intelligence tools created by Anthropic, the BBC reported on Monday. A department official told the BBC that the Pentagon has fully ended its use of products developed by the San Francisco-based firm. [1][2][7]

According to the BBC, Defence Secretary Pete Hegseth labelled Anthropic a supply chain risk in February after the company refused to remove safety guardrails from its tools. Secretary Hegseth had initially stated that the Pentagon would stop using Anthropic software by late August, although the exact reason for the delay until October was not immediately clear. [1]

The conflict has also involved public criticism from defence officials toward Anthropic leadership. Secretary Hegseth publicly criticised Dario Amodei, the chief executive of Anthropic, following the dispute over the removal of safety controls on the software. [1]

Anthropic was founded in January 2021 by former OpenAI members, including siblings Daniela Amodei and Dario Amodei, with an emphasis on artificial intelligence safety. The company creates the Claude series of large language models and reached a valuation of 965 billion dollars in a May 2026 funding round. [7]

Coverage from FourWeekMBA, AOL News, and The Times of India confirmed the official statement provided to the BBC. The decision marks an absolute end to the defence agency's operational reliance on Anthropic models. [3][4][5]