U.S. AI company Anthropic said in a misuse report published in September that a cell operating in Houthi-controlled northern Yemen had used its Claude model in three weapons development programs. The report returned to the headlines after Britain’s Daily Mail covered it on Oct. 12.
The report did not name the group explicitly, referring only to “a cell of threat actors based in northern Yemen.”
Guided rockets and a hypersonic goal

According to the company, the programs included a guided rocket with terminal homing, a multi-stage ballistic missile with a range goal above 2,000 kilometers and a missile with a hypersonic variant. The group test-fired a guided rocket, which appeared to fail, and returned to Claude within hours to analyze why.
Anthropic stressed there was no evidence the group had fielded a working weapon and said it banned the relevant accounts as soon as the activity was detected. Figures circulating recently, such as a 2,400-kilometer range and Saudi targets, do not appear in the company’s report.
The news comes as the Houthis continue missile and drone attacks on targets in Saudi Arabia, most recently striking King Khalid International Airport in Riyadh on Oct. 10.