Pentagon stops using Anthropic AI tools after blacklisting company, BBC told

According to a report by the BBC, the United States Department of Defense has ceased the use of AI tools developed by the startup Anthropic. The Pentagon reportedly blacklisted the company in February, citing concerns over what it described as a 'supply chain risk.' This decision followed a standoff between the defense agency and the AI firm regarding the implementation of safety guardrails.

Sources indicate that the conflict arose when Anthropic declined to remove specific safety protocols embedded within its artificial intelligence models. The Pentagon, which maintains strict requirements for the integration of third-party software into its secure systems, deemed the presence of these non-negotiable guardrails to be incompatible with its operational security standards. The move highlights the growing tension between the rapid adoption of generative AI in government sectors and the stringent security and control requirements mandated by military and intelligence agencies.

Anthropic, a prominent player in the AI industry known for its focus on 'constitutional AI' and safety-first development, has not publicly commented on the specific details of the Pentagon's internal security assessment. The suspension underscores the broader challenges facing defense departments globally as they navigate the integration of advanced machine learning tools while balancing proprietary software constraints with national security protocols. The Pentagon has not yet announced a replacement for the services previously provided by the firm.