Ukrainian units are increasingly deploying computer-vision systems to guide drones, easing operator workload while keeping humans in control.
As Bild reports, after analyzing strike videos published by the Azov 1st Corps of the National Guard, the use of one-way attack drones equipped with automatic target-detection systems is gradually turning into routine practice for Ukraine’s military.
The footage, shot in the Dobropillia area near Pokrovsk, shows a new level of technological support for drone operators, the outlet notes.
The videos display a drone interface with automatic object detection: potential targets are boxed in green; pre-recognized targets appear in purple. After the operator approves an attack with the cursor, the target is highlighted in red.
In other words, a human makes the final decision to strike; the system helps structure the visual feed.
Journalists note the system can identify and track multiple targets simultaneously. The operator can set priorities—for example, selecting the most important objects to attack.
This approach reduces the operator’s cognitive load but keeps them in the decision-making loop.
Sergey Sumlenny, director of the European Resilience Initiative Center in Berlin, told BILD that fully automatic strike missions are not being used at this stage. The front line remains too chaotic, he said, creating a risk of hitting friendly equipment.
That is why AI systems are used as assistive tools rather than autonomous decision-makers.
BILD says it received confirmation of the footage’s authenticity from a Ukrainian source involved in the practical deployment of drones. The source confirmed that the videos indeed use “real AI—insofar as one can call it that.”
He added that specialists prefer the more precise term “advanced computer vision.”
