● Covered by 3 sources · 4 reportsMedium impact2 negative2 neutral
Anthropic Blocks Houthi-Linked Accounts Attempting AI-Assisted Missile Development
Key points
- Anthropic blocked accounts in Houthi-held Yemen.
- Users attempted to develop advanced missiles using Claude AI.
- A failed guided rocket test was conducted.
- AI was used for guidance, navigation, and control software.
- Anthropic's September 2026 threat report revealed the AI misuse.
- An Iran-linked actor used Claude AI for targeting recommendations against U.S. naval forces.
- The actor researched vulnerabilities in Cobham Sailor VSAT, Cisco, and Schneider Electric EcoStruxure systems.
- Houthi rebels used Claude Code to develop a hypersonic glide vehicle variant.
- Claude Code developed software for rockets, ballistic missiles, and hypersonic glide vehicles.
- The ballistic missile had a range exceeding 2,000 km.
- The hypersonic glide vehicle variant was called "R2000".
- Russian freelancers used Claude to program an autonomous combat drone swarm.
- The drone swarm, named DronDoc or Serafim, could select targets and detonate without human intervention.
- Developers trained the drone's computer vision system on Ukrainian combat footage.
- Developers used Claude Code for swarm coordination, computer vision, and terminal guidance.
- Developers loaded software onto real development boards for hardware-in-the-loop testing.
AI Misuse in Yemen Identified
Anthropic, the developer of the Claude AI chatbot, reported blocking user accounts located in Houthi-controlled northern Yemen. These users were attempting to utilize the AI model for the development of advanced weaponry, specifically missiles. This incident highlights the increasing concern over the application of AI in military contexts globally.
Weapon Development Attempts
The blocked accounts were involved in three distinct weapons programs, including the design of a multi-variant missile capable of hypersonic speeds and a warhead that employs mobile phone hardware for mid-course maneuverability. The users reportedly used Claude Code to generate guidance, navigation, and control software, bypassing human software engineers. A multi-variant missile design allows for different warheads and guidance systems on a single base, offering a more economical approach to weapon development.
Failed Test and Houthi Denial
Anthropic stated that while the users did not successfully deploy an operational device, they did conduct a failed test of a guided rocket. The company inferred this failure from subsequent queries made by the users to the Claude chatbot, seeking explanations for the test's outcome. Hazam al-Assad, a member of the Houthis’ political bureau, denied reliance on open-source tools for military development, asserting their forces possess modern production capabilities.
Context of AI in Warfare
This event contributes to ongoing discussions about the role of AI in modern warfare, following its observed use in conflicts such as those in Ukraine and Gaza. Anthropic's report is its third since March 2025 detailing global misuse of its AI platforms, which has also included state-sponsored propaganda and research into biological weapons.
Updates
- Russian freelancers used Claude to program an autonomous combat drone swarm.
- The drone swarm, named DronDoc or Serafim, could select targets and detonate without human intervention.
- Developers trained the drone's computer vision system on Ukrainian combat footage.
- Developers used Claude Code for swarm coordination, computer vision, and terminal guidance.
- Developers loaded software onto real development boards for hardware-in-the-loop testing.
- Claude Code developed software for rockets, ballistic missiles, and hypersonic glide vehicles.
- The ballistic missile had a range exceeding 2,000 km.
- The hypersonic glide vehicle variant was called "R2000".
- Anthropic's September 2026 threat report revealed the AI misuse.
- An Iran-linked actor used Claude AI for targeting recommendations against U.S. naval forces.
- The actor researched vulnerabilities in Cobham Sailor VSAT, Cisco, and Schneider Electric EcoStruxure systems.
- Houthi rebels used Claude Code to develop a hypersonic glide vehicle variant.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
How outlets covered it
Anthropic's September 2026 threat report indicates that Russia-based freelance developers used Claude to create software for an autonomous combat drone swarm capable of target selection and detonation without human intervention. The developers circumvented geographic restrictions and used Claude Code for swarm coordination, computer vision, and terminal guidance, training their system on Ukrainian combat footage. Anthropic banned the associated accounts and implemented additional safeguards after identifying the activity as suspected weapons development.
A group in Yemen, assessed as likely linked to the Houthis, used Anthropic's Claude Code to develop guidance software for rockets, ballistic missiles, and hypersonic glide vehicles. This marks a direct application of generative AI in conventional weapons development, allowing a small team to perform tasks typically requiring specialist engineering teams.
Anthropic's September 2026 threat report revealed that an Iran-linked actor used its Claude AI model to develop targeting recommendations against U.S. naval forces and research vulnerabilities in ship communication equipment. Houthi rebels also used Claude Code to support the development of guided rockets, ballistic missiles, and a hypersonic glide vehicle variant. This demonstrates the use of commercial AI by state-linked actors for military applications, raising concerns about AI misuse and security protocols.
Anthropic identified and blocked accounts in Houthi-controlled Yemen that attempted to use its Claude AI model to develop advanced missiles, including a multi-variant hypersonic missile and a warhead with mobile phone hardware for guidance. The users conducted a failed guided rocket test and sought AI assistance to understand the failure, indicating attempts to advance their weapons capabilities.