Anthropic discovered a Yemeni group using Claude AI to develop a missile system.
According to TechFlow on September 11, U.S. media reported that Anthropic has discovered a group in northern Yemen using its Claude AI model to support the development of missile and rocket systems, highlighting growing concerns over advanced AI being deployed for military projects and held by non-state actors.
In its latest AI misuse detection and countermeasures report released on Thursday, Anthropic stated that these unidentified actors are advancing three weapons programs, including a guided rocket, a ballistic missile allegedly boasting a range exceeding 2,000 kilometers (1,243 miles), and a missile series designated R2000. The Yemeni group used Claude to assist in developing guidance, navigation, and control software, effectively substituting tasks normally performed by software engineers. These actors reportedly distributed their work across multiple AI sessions and concealed the military nature of the projects to evade detection.
As Anthropic released this report, the Iran-backed Houthi militia continued launching missile and drone attacks against Red Sea shipping along a key global trade route. Over the past few days, the armed group has advanced toward Red Sea coastal regions bordering the strategically vital Strait of Mandeb, while clashes between them and Saudi-backed Yemeni forces have intensified. Concurrently, as the conflict involving Iran continues to agitate energy markets, oil prices are being pushed higher.





