US weighs cutting ties with Anthropic over AI usage dispute

Amid growing tensions over how its AI models can be deployed, the U.S. Department of Defense is considering ending its partnership with Anthropic, highlighting broader ethical and strategic rifts between military requirements and AI safety guardrails.

The U.S. Department of Defense (DoD) is reportedly weighing the possibility of terminating its relationship with artificial intelligence company Anthropic following prolonged disagreements over how its AI technologies can be used by the military. The talks, which have involved the Pentagon and several leading AI developers, have stalled amid conflicting views on permissible deployments.

According to multiple media reports, Pentagon officials have been urging front-rank AI firms — including OpenAI, Google and xAI — to agree to terms allowing their technologies to be used for “all lawful purposes,” encompassing weapons development, intelligence analysis and battlefield operations. While other companies have shown flexibility in negotiations, Anthropic has resisted broad usage terms, citing ethical boundaries it has placed on its models.

Clash over ethical guardrails and military applications

Anthropic, known for its safety-focused AI systems such as the Claude model, has emphasised restrictions against deployment in fully autonomous weapons systems and large-scale domestic surveillance. The company has stated that its discussions with U.S. authorities have centred on these usage policy issues rather than specific battlefield operations. It also maintains a commitment to supporting national security objectives while advocating for robust ethical safeguards.

The impasse has angered some senior Pentagon officials, who see the limitations as a barrier to integrating advanced AI across military platforms and classified networks. As a result, the Defence Department is reportedly examining options that could include ending Anthropic’s contract — reported to be worth up to $200 million — or even designating the firm as a “supply chain risk,” a serious classification that could compel defence contractors to sever their ties with the company.

Broader AI and national security implications

The dispute comes against the backdrop of broader Pentagon efforts to accelerate AI adoption across defence planning, operations and intelligence systems, with a push for clearer, unified access to AI tools. Anthropic’s stance highlights the wider tension between rapid military integration of AI technologies and industry-led ethical considerations, signalling a pivotal moment in how artificial intelligence is governed and deployed in national security contexts.