Ukraine Expands AI Drone Warfare as Manufacturers Improve Targeting

Ukrainian defense manufacturers are equipping first-person-view drones with artificial intelligence that can recognize targets, maintain tracking and guide aircraft during the final stage of an attack. The technology is intended to reduce operators' workloads and help drones function when GPS signals or communications links are disrupted, as Ukraine and Russia expand their use of increasingly autonomous weapons.

NPR reported on October 29 that Ukrainian companies The Fourth Law and Odd Systems are producing AI-powered drones, including the Lupynis-10-TFL-1, which dozens of Ukrainian military units already use. Separately, United24 Media reported on October 28 that TAF Industries had introduced the Kolibri 10-inch Advance, incorporating an AI targeting module developed by DOD Solution.

The Lupynis-10-TFL-1 combines an affordable FPV drone with an onboard computer that assumes terminal guidance during the final 500 meters of flight. The system uses image recognition and a deep-learning neural network trained on hours of combat footage to identify objects such as vehicles, lightly armored vehicles, tanks and aircraft.

During a test observed by NPR at a site outside Kyiv, an operator using the military call sign Professor selected a target and monitored the drone's video feed. After the system locked onto a car moving along a dirt road, Professor released the controls and the drone guided itself toward the vehicle. The test aircraft carried no explosives.

Yaroslav Azhnyuk, a computer scientist who leads The Fourth Law and Odd Systems, said AI could help Ukraine compensate for Russia's greater military resources.

"When the other side has orders of magnitudes higher capabilities, it gives you an asymmetric advantage," he said.

Azhnyuk described the potential increase in operational effectiveness.

"It can make you 10,000 times more effective," he said.

Before Russia's full-scale invasion in 2022, Azhnyuk had worked in Silicon Valley and founded companies, including Petcube, which makes AI-enabled cameras for monitoring pets. He later redirected his work toward military technology.

Explaining how he viewed the risks associated with destructive technologies, Azhnyuk said concerns about their misuse predated artificial intelligence.

"It has been with humanity since the invention of fire or steel," he said.

He emphasized the role of human decisions in determining how technologies are used.

"Any kind of tool can be used for good or for bad. And in the end, it's the human that decides how to use that tool," he said.

Azhnyuk also described the contrast between his earlier consumer products and his current defense work.

"These cameras allow pet parents to watch pets when they're not at home," he said.

He explained that the products included interactive features for animals.

"and fling treats to them and shine a laser toy for them. We sold a million of these devices that are quite popular in the U.S.," he said.

Russia's invasion prompted his shift into defense production.

"I was like, OK, I've got to move on from Petcube and do something to protect my country," he said.

He described the new purpose of the camera technology used in his company's drones.

"And so now we're making not cameras that fling treats but cameras that fling explosives at people who came to occupy our land," he said.

Azhnyuk said his companies focus on autonomous drones and thermal imaging equipment.

"We are leaders in autonomous drones and thermal cameras," he said.

He distinguished the drones' onboard systems from larger AI applications.

"But it's very different from, say, the kind of AI that sits on those data centers that power Claude," he said.

He also contrasted the technology with AI used in self-driving vehicles.

"Or even the kind of AI that does self-driving for your Tesla," he said.

The systems use image recognition to support navigation and target tracking.

"which then allows you either to navigate autonomously without a human or to choose a target autonomously or to fly toward the target autonomously," Azhnyuk said.

The Lupynis drone takes its name from Ukrainian poet and political activist Anatoliy Lupynis, who endured decades of Soviet imprisonment and psychiatric torture. The company said it names its weapons after Ukrainian intellectuals and artists killed or persecuted by Russians during the previous century.

Azhnyuk explained the significance of the naming practice.

"It is important for us to help to resurrect those people," he said.

He connected the names to the continued influence of those figures.

"Their ideas are still with us. And now they're fighting back," he said.

During the field demonstration, a member of The Fourth Law's test and integration team, identified by the call sign Stab, pointed out the camera and onboard computer.

"So you see here is our camera, plus the microcomputer inside this 3-D model," Stab said.

He described the computer as the core of the drone's autonomous functions.

"All of the mind is here," he said.

Stab said the system can recognize several categories of potential targets.

"Like vehicles, light-armored vehicles, tanks, and flying objects," he said.

When the system identified the car during the test, Stab explained the indicator shown to the team.

"That means it has locked," he said.

As Professor released the controls, the drone continued toward the vehicle without further manual steering.

"See, Professor is doing nothing, and the drone is doing everything itself," Stab said.

Stab said AI-assisted guidance could make drones harder to disrupt through electronic warfare. He also said safeguards were intended to prevent The Fourth Law's drones from operating outside their intended role, emphasizing that humans select targets.

Azhnyuk rejected comparisons between current systems and fictional machines that independently turn against people.

"They're very far from a Terminator kind of scenario, right?" he said.

He compared their capabilities with ordinary computational tools.

"These are just sort of calculators. Like they're no more likely for a Terminator uprising than a warehouse with artillery shells," he said.

However, he acknowledged that future developments could bring more autonomous systems.

"It's not unthinkable," he said.

Ukrainian President Volodymyr Zelenskyy has warned that AI could influence battlefield decisions as soon as the following year, according to NPR's reporting. The expanding use of autonomous targeting has raised concerns about how much control people will retain over weapons decisions.

TAF Industries' Kolibri 10-inch Advance uses AURA AI, a module developed by DOD Solution and installed on a custom single-board computer designed for onboard AI processing. According to United24 Media, the system can identify moving and stationary targets, highlight potential threats for the pilot and continue terminal guidance after the operator confirms an engagement.

Andrii Atamanenko, TAF Industries' product owner for FPV drones, described the module's role in identifying possible targets.

"The automatic terminal-guidance module from DOD Solution recognizes targets, increasing the drone’s level of autonomy using AI. This function primarily helps pilots see all potential threats and targets by outlining them with a rectangle. Essentially, the AI module highlights targets the pilot might have missed," Atamanenko said.

United24 Media reported that the Kolibri 10-inch Advance had already been used by Ukraine's 93rd Separate Mechanized Brigade and 3rd Separate Assault Brigade. TAF Industries said integrating AURA AI into a new drone platform typically takes about one to two weeks, although the Kolibri project required additional preparation and testing.

Earlier reports cited by United24 Media indicated that successful strikes on Russian targets using AI-powered automatic homing technology had increased tenfold since the beginning of the year.