AI weaponized: Houthi rebels used Claude to target US ships and build missiles
The cutting edge of artificial intelligence is no longer confined to tech labs; it is rapidly becoming a pivotal, and deeply troubling, component in global geopolitical conflict. While some observers point to spiritual leaders using colorful rhetoric against the United States, a more stark reality emerges when examining how advanced AI tools are being deployed by state and non-state actors to enhance military capabilities, surveillance, and propaganda.
A recent threat intelligence report from Anthropic highlights a chilling application of this technology. It reveals that actors linked to Iran, alongside groups like the Houthis, are actively leveraging American-built AI models, such as Claude, to develop sophisticated offensive and defensive strategies against U.S. interests in the Middle East.
One of the most alarming findings involves the use of this technology for military targeting. A threat actor connected to Iran reportedly utilized Claude to analyze publicly available data—including ship and aircraft transponder identifiers, satellite imagery, and even names extracted from military photographs—to formulate targeting recommendations against U.S. naval forces. This capability demonstrates a disturbing fusion of open-source data and powerful AI to pinpoint vulnerabilities.
Beyond targeting, the AI has been instrumental in accelerating weapons development. A cell operating in northern Yemen, controlled by the Houthis, utilized Claude Code to support the development of complex weaponry. This included creating guidance, navigation, and control software for a guided rocket, designing multistage ballistic missiles capable of reaching over 2,000 km, and developing variants of the R2000 missile family, including hypersonic glide vehicles.
Instead of requiring a large team of software engineers, these groups reportedly used multiple instances of Claude to handle the entire engineering pipeline—from coding and code review to parameter tuning and flight simulations. This demonstrates how rapidly AI can be integrated into high-level military research and development, speeding up the creation of advanced systems.
The application of AI extends into domestic surveillance and information warfare as well. Iran-linked security units have reportedly used Claude to analyze a massive volume of social media data, profiling and surveilling thousands of opposition individuals over a single year. Furthermore, other groups have engineered domestic tracking tools, such as the “al-Najm al-thāqib” Firefox extension, designed for mass harvesting of user identities across major social platforms.
This trend is not isolated. While Anthropic has taken action by banning accounts associated with Iranian paramilitary and domestic security agencies, the broader reality is that various entities are pursuing these capabilities. Nations like China and Russia are also actively integrating Claude technology into their own extensive military programs.
The emergence of AI as a tool for sophisticated, asymmetric warfare underscores a critical shift in global security. The capability to deploy advanced models for surveillance, targeting, and weapons design presents a profound challenge to international stability, demanding a serious reevaluation of how we secure the digital future.
Buy on Amazon
Credit
