{"slug": "anthropic-reports-russian-developers-used-ai-for-kamikaze-drone-software", "title": "Anthropic reports Russian developers used AI for kamikaze drone software", "summary": "Anthropic published a 154-page threat intelligence report on September 11, 2026, detailing how freelance Russian developers used its Claude Code AI coding assistant to build autonomous targeting and detonation software for kamikaze drones aimed at Ukraine, a project tracked as GTG-27005 and internally designated DronDoc or Serafim. According to Anthropic, the developers began work around mid-May 2026, trained a computer vision classifier on Ukrainian combat footage focused on the Donetsk region, and used VPNs and commercial servers to circumvent the company's geographic restrictions; Anthropic assessed the group as freelancers with connections to a Russian regional university and a federal research center, and noted no confirmed operational deployment of the drone system, which remained in simulation and testing phases.", "body_md": "Anthropic official brand assets (anthropic.com)\n\n# Anthropic reports Russian developers used AI for kamikaze drone software\n\nA 154-page threat intelligence report reveals freelance developers in Russia used Claude Code to build autonomous targeting systems for attack drones aimed at Ukraine\n\nAnthropic just published what might be the most unsettling AI safety document to date: a 154-page threat intelligence report detailing how a group of Russian developers used Claude Code to build software for autonomous kamikaze drones designed to identify targets and detonate without human input.\n\nThe project, internally designated DronDoc or Serafim and tracked by Anthropic as GTG-27005, was designed to give attack drones the ability to select targets, distinguish between friendly and enemy forces, and execute terminal guidance maneuvers, all culminating in an autonomous detonation command. No human in the loop for the final kill decision.\n\n## What the developers actually built\n\nAccording to Anthropic’s report, released on September 11, 2026, the Russian developers began work on the project around mid-May 2026. They leveraged Claude Code, Anthropic’s AI coding assistant, to develop a suite of capabilities that reads like a military contractor’s wish list.\n\nThe software incorporated a computer vision classifier trained on Ukrainian combat footage, with a particular focus on the Donetsk region. It could perform target selection, including the ability to differentiate between what it classified as “enemy” and “friendly” targets, and it featured terminal guidance systems for the final approach phase.\n\nPerhaps most concerning: the system was designed so that drones could decide to detonate autonomously. The developers used a fixed coordinate in Ukrainian Donetsk for demonstration strikes, specifically targeting simulated frontline areas and logistics positions.\n\nTo access Claude Code in the first place, the developers circumvented Anthropic’s geographic restrictions using VPNs and commercial servers.\n\n## Freelancers with interesting connections\n\nAnthropic assessed the developers as freelancers rather than state-sponsored actors. The report noted connections between the group and both a Russian regional university and a federal research center.\n\n## Part of a broader pattern\n\nThe drone software wasn’t an isolated incident. Anthropic’s report placed the DronDoc project within a larger pattern of Russian-linked operations that used Claude for cyber espionage activities. These parallel efforts targeted drone technology and related assets in Ukraine, with the cyber operations involving data theft and attacks against various Ukrainian entities.\n\nAnthropic did note one critical caveat: no confirmed operational deployment of the drone system has been documented. The project remained in simulation and testing phases.\n\n## The implications for AI companies and regulators\n\nAnthropic deserves some credit for transparency here. Publishing a 154-page report detailing how your own product was used to develop weapons software is not exactly a PR win. But it does set a precedent for how AI companies might handle disclosure when their tools get repurposed for military applications.\n\nThe computer vision component adds another layer of complexity. Training a targeting classifier on real combat footage suggests that publicly available conflict documentation, the kind shared by open-source intelligence communities and war correspondents, can become training data for weapons systems.\n\n**Disclosure:** This article was edited by Editorial Team. For more information on how we create and review content, see our\n\n[Editorial Policy](https://cryptobriefing.com/editorial-policy/).", "url": "https://wpnews.pro/news/anthropic-reports-russian-developers-used-ai-for-kamikaze-drone-software", "canonical_source": "https://cryptobriefing.com/anthropic-russian-ai-kamikaze-drone-software/", "published_at": "2026-09-12 14:40:36+00:00", "updated_at": "2026-09-12 14:45:35.945675+00:00", "lang": "en", "topics": ["ai-safety", "ai-policy", "computer-vision", "ai-tools", "ai-ethics"], "entities": ["Anthropic", "Claude Code", "Russia", "Ukraine", "Donetsk", "GTG-27005", "DronDoc", "Serafim"], "alternates": {"html": "https://wpnews.pro/news/anthropic-reports-russian-developers-used-ai-for-kamikaze-drone-software", "markdown": "https://wpnews.pro/news/anthropic-reports-russian-developers-used-ai-for-kamikaze-drone-software.md", "text": "https://wpnews.pro/news/anthropic-reports-russian-developers-used-ai-for-kamikaze-drone-software.txt", "jsonld": "https://wpnews.pro/news/anthropic-reports-russian-developers-used-ai-for-kamikaze-drone-software.jsonld"}}