US Military Pressures Anthropic to Remove Claude Safeguards for Military Use

The US Department of Defense is pressuring Anthropic to remove ethical safeguards from its Claude AI model to allow military applications that the company currently restricts. Defense Secretary Pete Hegseth met with Anthropic CEO Dario Amodei on Tuesday and gave the company until Friday to agree to the Pentagon's terms or face penalties.

Key Details from the Negotiations

According to the source material:

• The Pentagon wants "unfettered access" to Claude's capabilities for military use
• Anthropic has resisted allowing Claude to be used for mass surveillance or autonomous weapons systems that can kill without human input
• The DoD has threatened to cancel its contract with Anthropic and designate the company as a "supply chain risk" if it doesn't comply
• The DoD signed deals with several AI companies including Anthropic, Google, and OpenAI in July 2025 worth up to $200 million
• Until this week, Claude was the only AI model permitted in the military's classified systems
• The DoD signed a deal on Monday allowing use of Elon Musk's xAI chatbot in classified systems
• Both xAI and OpenAI have agreed to the government's terms, with OpenAI allowing its model to be used for "all lawful purposes"

Background and Context

The meeting follows reports that the US military used Claude to assist in the capture of Venezuelan leader Nicolás Maduro last month. There's been a push from the Trump administration to integrate AI into military operations, with Donald Trump vowing that the US will win a global AI arms race.

Pentagon Chief Technology Officer Emil Michael has publicly campaigned for Anthropic to "cross the Rubicon" and agree to the government's terms, stating: "I think if someone wants to make money from the government, from the US Department of War, those guardrails ought to be tuned for our use cases – so long as they're lawful."

Anthropic has positioned itself as the most safety-focused AI company and has backed political action committees advocating for stronger AI safeguards. The company has hired several former Biden staffers, which reportedly contributed to a pro-Trump venture capital firm backing out of investing in Anthropic earlier this year.

Broader Implications

The Pentagon has invested billions in AI technologies including unmanned drones and automated targeting systems. The ethical debate around AI decision-making for lethal force is no longer theoretical, with Ukraine featuring semiautonomous drones that can operate without human control.

This situation highlights the tension between AI companies' ethical guidelines and government demands for military applications. For developers working with AI agents, it raises questions about how model safeguards interact with different deployment environments and the legal frameworks governing AI use.