Understand
Pentagon stops using Anthropic AI tools after national security designation
Overview
The US Department of Defence has confirmed it has ceased all use of Anthropic's AI tools, months after Defence Secretary Pete Hegseth designated the company a national security supply chain risk. The designation followed Anthropic's refusal to remove safety guardrails from its Claude model to give the military unfettered access. Despite the ban and a February deadline, sources revealed that Claude remained integrated into the Pentagon's Palantir-operated Maven Smart System for intelligence gathering and research until recently. Anthropic has sued the Trump administration over the risk designation, though recent meetings have taken place between President Trump and Anthropic boss Dario Amodei.
Key points
- The US Department of Defence has officially confirmed that it has ceased all use of Anthropic's products.
- Defence Secretary Pete Hegseth designated Anthropic a national security supply chain risk in February after the company refused to remove safety guardrails from its Claude model.
- Sources familiar with the matter stated that Anthropic's tools were still being used for research, analysis, and intelligence gathering as recently as last week.
- Claude was deeply embedded in the Pentagon's primary intelligence organisation platform, the Palantir-operated Maven Smart System, making quick removal difficult.
- Anthropic has sued the Trump administration over the supply chain risk designation, while the Pentagon has formed contracts with other tech firms including OpenAI, xAI, and Google.
- Anthropic boss Dario Amodei recently met with President Donald Trump at the White House to discuss AI risks.