Anthropic blocks Yemen-based users who tried to build missiles with Claude AI

Anthropic has disclosed that users in northern Yemen attempted to leverage its Claude AI chatbot to develop advanced missile systems, according to a report the company released Thursday.

The AI developer identified and blocked accounts in territory controlled by Iran-backed Houthi rebels that were pursuing weapons programs including a multi-variant missile designed to glide at hypersonic speeds and a warhead using mobile phone hardware for mid-course maneuvering. The users employed Claude Code as a substitute for human software engineers to develop guidance, navigation, and control software.

While the actors did not succeed in fielding an operational device, Anthropic said they conducted a failed test of a guided rocket. The company learned of the failure because the users returned to Claude to troubleshoot why their test had not worked. Before their accounts were banned, the users had built an offline simulation toolkit that operates independently of Claude or other computing platforms.

Houthi denial and weapons capabilities

Hazam al-Assad, a member of the Houthis' political bureau, dismissed the report as "unreasonable and illogical," stating that the group's armed forces possess "modern, diverse and developed production capabilities" accumulated during Yemen's 12-year civil war. He characterized all weapons as defensive.

Weapons analyst Trevor Ball of Armament Research Services expressed skepticism about the Houthis' ability to actually produce hypersonic missiles, noting that even U.S. hypersonic weapons remain in testing. He suggested the group is likely attempting to reduce dependence on Iranian weapons shipments by developing indigenous capabilities. The Houthis already possess Iranian-made anti-ship missiles with mid-course guidance adjustment systems.

Broader AI misuse patterns

The Yemen findings appeared in Anthropic's third global misuse report since March 2025, covering the period from December to August. The report documented various malicious uses of AI platforms, from state-sponsored propaganda campaigns to unnamed actors researching methods to enhance biological weapons lethality.

Adam Baron, a Yemen researcher at the New America think tank, said the incident reflects the Houthis' growing technological sophistication. "There's a tendency to see the Houthis as this group of barefoot tribal fighters, and that's just not true," he said, pointing to their manipulation of social media narratives and ability to absorb expertise from Iran and allied groups.

Why it matters

The incident underscores how commercially available AI tools are becoming accessible to non-state armed groups in active conflict zones, potentially accelerating weapons development in regions where traditional manufacturing constraints might otherwise limit capabilities. As AI models grow more powerful and widespread, the challenge of preventing their misuse in warfare intensifies—even when, as in this case, the attempts ultimately fail. The disclosure also raises questions about how AI companies can effectively monitor and restrict use by actors in regions where attribution and enforcement are difficult.

Anthropic said it shared its findings with private and public partners. The Houthis have recently made significant territorial gains in Yemen and continue threatening key shipping routes in the Red Sea and Bab al-Mandeb Strait.

These details were first reported by the Associated Press.