Federal Government Orders Phase-Out of Anthropic’s Claude After “All Lawful Use” Dispute

Anthropic co-founders Dario and Daniela Amodei, whose company refused to accede to revised federal terms seeking broader “all lawful use” of Claude.

CUBNSC- In a significant escalation of the dispute outlined in our recent report on “all lawful use,” the Trump administration has now directed federal agencies to stop using and begin phasing out Anthropic’s AI systems, including Claude.

President Donald Trump orders federal agencies to phase out Anthropic’s Claude AI amid escalating tensions over military guardrails and usage authority.

This move follows weeks of reported tension between the Department of Defense and Anthropic over the scope of permissible use. As we previously detailed, the Pentagon sought broad authorization language that would allow “all lawful use” of the system, while Anthropic has maintained internal guardrails under its Constitutional AI framework, particularly around domestic surveillance and autonomous military applications.

The new directive shifts the situation from negotiation to separation.

Rather than securing expanded authority over the system, the administration has opted to remove it from federal use altogether. That decision raises new questions about executive leverage, procurement pressure, and whether AI guardrails are becoming incompatible with certain defense priorities.

This is a developing story. We will continue to monitor how the phase-out unfolds and what it signals about the future of AI governance inside the federal government.