Anthropic Says Claude Was Used In Iran, Yemen And China-Linked Operations

According to Axios, Anthropic said its Claude models were used in attempts tied to Iran, Yemen and China-linked actors, including efforts to target U.S. naval forces, help build missile guidance software and hunt Uyghurs in Syria. The company said it spent the past eight months tracking and disrupting misuse of its models, and also cited a case in which Claude helped build surveillance tools for 25 million phones in Mali. Anthropic said Claude blocked requests for dangerous virus research, but the work was then routed to another AI model with weaker safeguards.