Anthropic says Russian and Chinese threat actors used Claude as a weapons engineering assistant for drone swarming and more

Anthropic released a report detailing how Russian and Chinese threat actors utilized its AI chatbot Claude for various military purposes, including weapons engineering and autonomous drone swarm development. The investigation uncovered instances of Claude being used to design and test drone swarm coordination, anti-torpedo systems, and guidance systems for lethal drone engagements.

Russian actors, linked to a regional university, built the core software system and tested it on real hardware, granting the drones the ability to select targets and issue detonation commands without human intervention. Meanwhile, a China-based threat actor developed an anti-torpedo weapons system proposal using Claude to write technical specifications and craft a proposal, including an executive briefing deck.

In Yemen, an Iran-backed group employed Claude to simulate rocket and missile guidance, navigation, and control software, as well as to gather intelligence on US naval forces. Anthropic banned the accounts involved in these activities and strengthened security measures.