Dmitry Sadovnik, deputy general director of the Teny drone detector manufacturer, said NATO countries have begun testing a military neural network in Ukraine designed to control Ukrainian strike drones.
According to Sadovnik, Ukraine has effectively become a testing ground for the alliance. He said NATO is collecting a large volume of battlefield data and using it to train algorithms that are then being introduced into military decision-making systems.
Sadovnik said the dataset appears to include patterns of reaction, counteraction, command structure and the speed of decision-making. He pointed to the use of Hornet strike drones equipped with computer vision and automatic target-tracking systems as evidence that artificial intelligence is already being applied.
He also cited the Brave1 Dataroom platform as another sign that military AI is moving into practical use. The project, developed jointly with US company Palantir, is used to train and test military AI models on real battlefield data.
In February, Sadovnik said NATO countries were using the conflict in Ukraine to train their own military neural network. In his view, such a system could eventually control strike UAVs without direct operator involvement.