Yemen-based group used Claude AI in missile development, Anthropic says
 |
| Claude Anthropic logo. File photo. |
PARIS — September 12, 2026
Anthropic said a group operating in northern Yemen used its Claude artificial intelligence chatbot in efforts to develop guided rockets and missiles, including software for navigation and targeting. The company said it detected the activity and banned the accounts involved, but found no evidence that the group successfully deployed an operational weapon.
In a report on efforts to prevent illegal uses of its AI systems, Anthropic said it identified a cell in northern Yemen that was working on three weapons development programmes. The projects included a guided rocket, a multi-stage ballistic missile and a missile design involving a hypersonic glide vehicle.
According to Anthropic, one of the projects involved a guided rocket using a commercially available, phone-class computer for flight control and final-stage guidance. Another was described as a multi-stage ballistic missile with a stated target range of more than 2,000 kilometres, while a third included several missile designs and a hypersonic glide vehicle variant.
The company said Claude was used to help develop guidance, navigation and control software for the airborne weapons. Anthropic said the use of the chatbot could reduce the need for a larger team of engineers working on those tasks.
Anthropic said it had not found evidence that the group successfully fielded an operational weapon. However, the company said the group test-fired a guided rocket and that the test appeared to fail. The actors then returned to Claude within hours to seek help understanding the failure, according to the report.
Anthropic did not identify the group. The company said the activity was carried out by actors based in northern Yemen, an area largely controlled by the Houthi movement, but it did not publicly confirm that the Houthis were responsible.
The disclosure comes as Yemen's conflict remains closely linked to tensions around the Bab al-Mandab Strait, a major maritime route connecting the Red Sea with the Gulf of Aden and wider trade routes between Europe and Asia. Houthi forces have continued military activity in the region.
Anthropic also said it had disrupted attempts to use Claude for developing biological weapons and in a suspected Russia-linked cyber espionage campaign targeting Ukraine. The company further accused Chinese competitors of attempting to extract capabilities from its AI systems.
The report highlights growing concerns about the use of advanced AI in military and other harmful activities. While Anthropic said its safeguards detected and stopped the accounts involved in the Yemen case, the episode shows how AI tools can potentially assist technical work that previously required specialized expertise.
Anthropic researchers have separately warned that increasingly capable AI systems could create serious risks for humanity if their development and use are not properly controlled. The company has continued to emphasize restrictions on using its models for harmful or illegal activities.
