First AI‑Killed Soldiers Claim in Ukraine War Tests Red Line on Autonomous Weapons
A Ukrainian drone maker says fully autonomous AI 'Terminator' drones killed Russian soldiers near Bakhmut two years ago, with no human in the firing loop. If confirmed, the field test would mark one of the first known uses of autonomous weapons to take lives in a major war — raising uncomfortable questions for militaries, lawmakers and engineers far beyond Ukraine.
Lethal autonomy in warfare may no longer be theoretical. A Ukrainian drone manufacturer has claimed that, two years ago, a swarm of fully autonomous AI‑guided drones killed Russian soldiers near Bakhmut and Chasiv Yar without any human directly controlling their targeting, pushing one of the most sensitive red lines in modern warfare from the lab to the battlefield.
Drone‑maker Alexander Kokhanovskyy told a science outlet that around ten so‑called "Terminator" drones were used in a one‑off test near the contested front in eastern Ukraine. According to his account, the drones were programmed to identify and attack Russian targets on their own, resulting in the deaths of “a couple of soldiers” and the destruction of a truck. No independent verification of the specific incident has yet been made public, but the description aligns with Ukraine’s rapid, documented expansion of AI‑enabled battlefield systems since 2022.
Kokhanovskyy’s remarks, echoed in separate reporting in Spanish, cut directly into a debate that has largely been confined to policy papers and diplomatic conferences. Ukraine currently bans full autonomy at the final targeting stage, requiring a human operator to approve lethal strikes. The drone maker says that rule was relaxed for a single test and that Kyiv is now openly discussing whether to loosen restrictions on fully autonomous targeting as the war grinds into its fourth year.
For soldiers on the front line, the stakes are literal life and death. Allowing software to decide when to fire removes the human judgment that, however imperfect, has long been seen as a final safeguard in combat. For Russian units under constant attack from Ukrainian drones, the possibility that some of those systems could choose targets without a human operator makes trenches, vehicles, and logistics convoys feel even more exposed. For Ukrainian troops, the promise is fewer of their own operators having to hover over screens, manually piloting individual drones into Russian positions under fire.
Strategically, the claim points to a looming arms race that extends well beyond the Donbas. Militaries from the United States and China to Turkey and Israel are investing in loitering munitions, swarming drones and AI‑enabled targeting, but most governments publicly insist a human will stay “in the loop” for lethal decisions. A battlefield precedent for autonomous kills in Europe’s largest land war since 1945 will make it harder for others to maintain that line, especially if commanders believe autonomy offers a decisive edge against a better‑equipped opponent.
Ukraine’s reported test also exposes a gap between formal policy and battlefield innovation. Even as Kyiv publicly endorses constraints on autonomous weapons, its defense sector is under intense pressure to offset Russia’s advantage in manpower and artillery with cheaper, smarter systems. In that environment, the temptation to let software handle more of the kill chain is strong — and other countries observing Ukraine’s experience may quietly follow.
The memorable point is simple: once one side in a high‑intensity war proves that code can reliably pick and hit targets on its own, the question for others is no longer whether to adopt lethal autonomy, but how fast they can do it while still claiming control.
What happens next will be shaped less by battlefield technologists than by lawmakers and defense ministries. Signals to watch include whether Ukraine formally revises its rules governing autonomous targeting, how Russia adjusts its air defenses and electronic warfare against smarter drones, and whether upcoming United Nations talks on autonomous weapons move from abstract principles toward concrete bans or constraints. Defense budgets in NATO and beyond will reveal how quickly governments are willing to fund not just more drones, but more decisions made by machines.
Sources
- OSINT