Pentagon Halts Anthropic AI Use After Blacklisting the Company
The Defense Department cut off Anthropic's AI tools after labeling the firm a supply chain risk over its refusal to remove safety guardrails.
The Pentagon has stopped using artificial intelligence tools developed by Anthropic following a decision to blacklist the company, BBC News has learned. The Defense Department designated Anthropic a "supply chain risk" in February, a classification that effectively barred its products from military use.
The root of the dispute centers on Anthropic's refusal to strip safety guardrails from its AI systems at the government's request. Those guardrails are built-in restrictions designed to prevent the technology from generating harmful or dangerous outputs, and Anthropic declined to disable them for Pentagon use.
Read more Trump Creates AI Policy Taskforce Led by Intelligence Chief →
The standoff highlights a deepening tension between AI developers committed to safety-first development practices and government agencies seeking maximum operational flexibility from commercial AI tools. Anthropic, which makes the Claude family of AI assistants, has positioned itself publicly as a safety-focused lab, making the removal of those protections a line the company was unwilling to cross.
The blacklisting represents a significant contract and reputational development for Anthropic, which has been competing aggressively for lucrative federal and defense contracts alongside rivals such as OpenAI and Google. Whether the two sides can reach any accommodation remains unclear, and the Pentagon has not publicly detailed what specific capabilities it sought.
Continue reading at BBC News.
