{"id":111607,"date":"2026-09-16T11:32:59","date_gmt":"2026-09-16T15:32:59","guid":{"rendered":"https:\/\/brocku.ca\/brock-news\/?p=111607"},"modified":"2026-09-16T17:33:32","modified_gmt":"2026-09-16T21:33:32","slug":"opinion-john-heckman-discusses-how-a-russian-team-trained-a-kamikaze-drone-using-ai","status":"publish","type":"post","link":"https:\/\/brocku.ca\/brock-news\/2026\/09\/opinion-john-heckman-discusses-how-a-russian-team-trained-a-kamikaze-drone-using-ai\/","title":{"rendered":"Opinion: John Heckman discusses how a Russian team trained a &#8216;kamikaze&#8217; drone using AI"},"content":{"rendered":"<p><em>This piece written by John Heckman, a PhD in Interdisciplinary Humanities candidate at Brock University<\/em><em>, originally appeared in <a href=\"https:\/\/theconversation.com\/russian-team-misused-claude-ai-to-train-a-killer-drone-on-scraped-ukrainian-war-footage-291965?\" target=\"_blank\" rel=\"noopener\">The Conversation<\/a>.<\/em><\/p>\n<p>A new\u00a0threat intelligence report from the AI company Anthropic\u00a0has revealed that a Russian team scraped Ukrainian combat videos to train an autonomous \u201ckamikaze\u201d drone to choose its own targets and destroy them.<\/p>\n<p>Anthropic\u2019s report was published on Sept. 10, two days before CEO and co-founder of the company, Dario Amodei, published\u00a0an essay arguing that humanity must urgently \u201cslow the pace\u201d of AI development.<\/p>\n<p>The report is a 154-page account of the attempted misuse of Claude AI models for \u201cmalicious activity\u201d between December 2025 and August 2026. It covers seven areas: cyber operations, influence operations, surveillance, scams and fraud, biological misuse, conventional weapons development and distillation.<\/p>\n<p>One case, tracked as GTG-27005, describes a small, likely freelance team in Russia that used Claude\u2019s coding tool to help build software for a swarm of\u00a0autonomous first-person-view (FPV) \u201ckamikaze\u201d drones\u00a0\u2014 the cheap, camera-guided munitions that have become one of the defining weapons of the war.<\/p>\n<p>This should worry anyone who has ever filmed, shared or trusted footage from the war in Ukraine \u2014\u00a0one of the most visually documented conflicts in history.<\/p>\n<p>It seems that any visual footage \u2014 even that created by journalists, soldiers, civilians and researchers to update the outside world and hold nations accountable \u2014 can now feed back into a conflict as raw material for weapons development and deployment on the battlefield.<\/p>\n<p><strong>Autonomous drone, can target a person<\/strong><\/p>\n<p>Anthropic\u2019s report states the company\u2019s threat intelligence teams have \u201cinvestigated and disrupted multiple threat actors\u201d who used Claude models to develop software for either weapons design and development or for intelligence-gathering and procurement for weapons programs.<\/p>\n<p>The report documents six cases: three in China, two in Russia and one in Yemen.<\/p>\n<p>The \u201clikely freelance Russia-based\u201d developers in case GTG-27005 called their project \u201cDronDoc\u201d or \u201cSerafim.\u201d They began work in May 2026 and routed their traffic through commercial servers to get around Anthropic\u2019s location-based restrictions.<\/p>\n<p>The software they built could enable a drone to pick its own targets, including targets in the category \u201cperson,\u201d and issue the command to detonate \u2014 without a human confirming the strike first. That detail made\u00a0headlines\u00a0and it\u00a0deserved the publicity.<\/p>\n<p>However, one line in the report has received far less attention: the developers trained their targeting software on scraped Ukrainian combat footage, sorted into \u201cenemy\u201d and \u201cfriendly,\u201d and tested it against a simulated target location in the region of the\u00a0Donetsk oblast.<\/p>\n<p><strong>Scraped Ukrainian combat footage<\/strong><\/p>\n<p>The report does not specify where the footage came from. It could have come from Telegram channels used by military personnel, official and unofficial drone strike compilation accounts, open-source intelligence (OSINT) archives or open databases assembled by journalists and researchers. At this time, we simply don\u2019t know which is the likely source.<\/p>\n<p>The\u00a0Centre for Information Resilience\u2019s\u00a0Eyes on Russia Map, for example, has verified and mapped more than 30,000 individual incidents using footage from volunteers, soldiers and civilians.<\/p>\n<p>Drone footage and the aftermath of missile strikes are verified and geolocated by researchers who hope that they\u2019ll be used as evidence for journalistic endeavours, war crimes investigations or simply to show the reality of the conflict in an era when generative AI can skew the narrative.<\/p>\n<p>Those producing this footage are, for the most part, trying to make the war more visible and those responsible more accountable for their actions.<\/p>\n<p><strong>Documented for evidence, used for conflict<\/strong><\/p>\n<p>What this case shows is that footage built for one purpose can be repackaged for a very different one without changing at all. A clip filmed to document a strike and one that is scraped to teach a weapon what a human target looks like are, frame for frame, the same file.<\/p>\n<p>Nothing about the image itself signals what use it will be put to. What determines that is who can access it and what they choose to do with it.<\/p>\n<p>Ukraine\u2019s own\u00a0drone video system, OCHI,\u00a0has amassed well over two million hours of footage since 2022, making it the largest publicly known collection of live-fire combat video globally.<\/p>\n<p>A dataset of that scale is a gold mine for machine learning. However, this is not unusual when it comes to modern, digital warfare. A company in the United States, Enabled Intelligence, states that they recently packaged more than\u00a0half a million hours of Ukrainian drone footage\u00a0into a commercial dataset which was sold for military AI training.<\/p>\n<p>Indeed, this kind of extraction is becoming routine and openly discussed. GTG-27005 demonstrates that it isn\u2019t confined to declared, sanctioned users.<\/p>\n<p><strong>A reverse-engineered vision system<\/strong><\/p>\n<p>This isn\u2019t an isolated case, either. The same report describes a separate incident, tracked as GTG-20006, and connected by Anthropic to the\u00a0Russian state-linked group publicly known as Midnight Blizzard.<\/p>\n<p>In that case, operators stole a software kit for a military drone\u2019s vision system and spent days using Claude to reverse-engineer it \u2014 recovering the design, components and details of an unreleased product.<\/p>\n<p>Two separate groups within the same period went after the same kind of target: the system that allows a drone to recognize what it sees. Is this the beginning of a pattern \u2014 one where a drone\u2019s ability to visually identify a target is becoming just as valuable as its engine or explosive payload?<\/p>\n<p><strong>Who else is watching?<\/strong><\/p>\n<p>None of this is an argument against documenting conflict. Far from it. Those visuals remain one of the few tools that ordinary people have to\u00a0hold power accountable during moments of strife.<\/p>\n<p>However, it is a reason to take seriously that those producing this record \u2014 journalists, soldiers, civilians volunteer researchers \u2014 are part of a supply chain whose subsequent uses they cannot predict and never agreed to.<\/p>\n<p>Most conversations about the ethics of war footage have focused on\u00a0who should watch\u00a0it\u00a0and how. This case suggests that we also need to ask who else is watching and what are they going to do with it.<br \/>\n<img decoding=\"async\" loading=\"lazy\" style=\"border: none !important; box-shadow: none !important; margin: 0 !important; max-height: 1px !important; max-width: 1px !important; min-height: 1px !important; min-width: 1px !important; opacity: 0 !important; outline: none !important; padding: 0 !important;\" src=\"https:\/\/counter.theconversation.com\/content\/291965\/count.gif?distributor=republish-lightbox-advanced\" alt=\"The Conversation\" width=\"1\" height=\"1\" \/><\/p>\n","protected":false},"excerpt":{"rendered":"<p>John Heckman, a PhD in Interdisciplinary Humanities candidate at Brock University, recently published a piece in The Conversation about how a Russian team trained a \u201ckamikaze\u201d drone using Claude AI and Ukrainian combat footage. <\/p>\n","protected":false},"author":1,"featured_media":111612,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[188,55,37,6],"tags":[384,4037,15257,5512],"_links":{"self":[{"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/posts\/111607"}],"collection":[{"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/comments?post=111607"}],"version-history":[{"count":5,"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/posts\/111607\/revisions"}],"predecessor-version":[{"id":111613,"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/posts\/111607\/revisions\/111613"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/media\/111612"}],"wp:attachment":[{"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/media?parent=111607"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/categories?post=111607"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/brocku.ca\/brock-news\/wp-json\/wp\/v2\/tags?post=111607"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}