Kyiv Drone Maker Tests AI Targeting in Overgrown Field

Yaroslav Azhnyuk stands in an overgrown field outside Kyiv, watching a small drone hunt a target.

The 37-year-old computer scientist runs The Fourth Law, a Ukrainian defense firm that manufactures AI-powered drones. His company also operates under the name Odd Systems. Dozens of Ukrainian military units currently fly their autonomous model, the Lupynis-10-TFL-1. Azhnyuk argues that artificial intelligence gives a smaller force a major advantage. He told NPR that AI can make a unit 10,000 times more effective when facing an opponent with superior capability.

The technology works by combining a standard first-person-view drone with an onboard computer. This unit takes control during the last 500 meters of flight. That distance equals 0.3 miles. The system uses deep-learning neural networks trained on hours of combat footage. It identifies vehicles, light-armored trucks, tanks, and flying objects. Stab, the head of the test and integration team, explained the process during a demonstration. He pointed to the microcomputer inside the drone, calling it the mind. The drone locks onto a target, which triggers a red flash on the pilot’s monitor. Once locked, the drone flies itself to the target without pilot input.

Azhnyuk transitioned to drone warfare after Russia launched its full-scale invasion in 2022. He previously lived in the United States and built Petcube, a company that sold one million smartphone-powered pet cameras. Those cameras used AI to watch pets and dispense treats. Azhnyuk described his shift in focus with dark humor. He said he moved from making cameras that fling treats to making cameras that fling explosives at occupiers.

The company names its weapons after Ukrainian intellectuals persecuted by Russia in the 20th century. The Lupynis drone honors Anatoliy Lupynis, a poet and political activist who endured Soviet imprisonment. Azhnyuk said naming the weapons helps resurrect those figures. He noted that their ideas are still present and now fighting back. The Fourth Law tests these drones daily in a field near Kyiv. Security concerns keep the exact location and team identities private. The team includes a pilot known by the call sign Professor. He guides the drone until the AI takes over for the final strike.

Ukrainian President Volodymyr Zelenskyy has warned that AI could guide battlefield decisions as soon as next year. Russia is also deploying AI-enabled drones against Ukraine. This development is part of a high-tech arms race that worries many leaders. Azhnyuk acknowledges the concerns about destructive weapons. He compares the anxiety over AI to the historical fears surrounding fire or steel. He believes humans ultimately decide how to use any tool. He insists his company’s drones are not killer robots. He described them as calculators, noting they are no more likely to rise up than a warehouse full of artillery shells.

Safeguards prevent the drones from going rogue. Humans select the target before the AI engages. Stab noted that AI drones are useful because they are hard to jam. Azhnyuk said his company is focused on current needs rather than future ethics. He argued that if an adversary uses AI for offensive decisions, refusing to use AI for defense leads to defeat. He said enemy systems can make decisions in microseconds, while human control takes minutes or hours. If Ukraine loses, he said, there is no way to argue about the ethics of AI. The priority, he said, is survival.