Drop Site
🇾🇪 ⭕️ Anthropic also says a Yemen-based weapons cell used its AI to write missile guidance software In its same Sept. 10 threat report, the company said actors in northern Yemen (likely Ansarallah, but report does not clarify) ran three weapons programs through Claude Code, using it in place of human software engineers. Per the report: ➤ A guided rocket with final-phase homing was field-tested in Yemen. The test appears to have failed, and the actors returned to the model within hours to diagnose why ➤ A ballistic missile simulation carried a stated range goal above 2,000 km ➤ A multi-variant set called the R2000 included a hypersonic glide vehicle The cell ran several model instances at once, assigning one to write code, another to research and a third to review, the report said. Anthropic said its safeguards blocked many requests but not all, and the actors split work across sessions to hide the full program. 👉 It banned the accounts, but said not before the actors had already built an offline simulation toolkit that runs without Claude.