Anthropic Disrupted Russian Attempt to Develop Autonomous Drone Swarms
Anthropic, the AI safety and research company said that it had prevented Russian developers from using its Claude coding agent to create software to power the Kremlin’s autonomous kamikaze drones. The system was designed to select and attack targets, including individuals, without the need for operator intervention.
Anthropic, an AI safety and research company, has uncovered a Russia-linked group developing software for autonomous kamikaze drones using its Claude coding agent. The group, named GTG-27005, aimed to create a swarm of autonomous combat drones capable of making decisions independently. This marks a new direction for the group, which previously focused on cyberattacks.
The team, believed to be independent university specialists, registered their accounts in late 2025 and intensified their operations in mid-2026. They used combat footage from Ukraine's conflict, obtained through social media, to train a computer vision system to identify hostile targets. The developers also created an onboard language model to determine each drone's next action following an engagement.
Anthropic identified and blocked nine accounts associated with this project, along with connections to the Russian Academy of Sciences and a regional university. The report highlights the need for oversight in AI development to prevent potentially catastrophic scenarios.
Written by urgent.news from Kyiv Post's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
