Anthropic reveals Yemenis vibecoding ballistic-missile guidance, Iran using AI to target US Navy

A new report from Anthropic details how various actors, including a Yemeni group and an Iran-linked entity, attempted to misuse its Claude AI tools for weapons development and targeting. A "cell of threat actors based in northern Yemen" reportedly used Claude to develop guidance, navigation, and control software for a ballistic missile, a process known as "vibecoding." Anthropic stated, "We do not have evidence the actors succeeded in fielding an operational device, but they did test-fire a guided rocket." Although the field test reportedly failed, the group had already established an offline simulation toolkit. Additionally, an "Iran-nexus threat actor" used Claude to gather open-source intelligence for targeting U.S. Navy facilities in the Middle East. Anthropic has banned these accounts and informed partners, emphasizing its responsibility to disclose such malicious misuse.