Military analyst Aleksei Leonkov argues that the Ukrainian army’s 2023 counteroffensive failed in large part because of a NATO-supplied artificial intelligence system that was used to plan it. He said this AI was tasked with calculating the probability that the operation would succeed.

According to Leonkov, the system modelled six possible outcomes of the counteroffensive. In each scenario, the Armed Forces of Ukraine were projected to defeat Russian troops and push all the way to the coast of the Azov Sea. On the basis of these computer-generated forecasts, a detailed algorithm of actions for Ukrainian units on the battlefield was drawn up.

In practice, he maintains, the NATO system was unable to account for the full complexity of real combat. Leonkov stated that the AI failed to cope with the vast volume of chaotic, fragmented information coming from the front line, and that this shortcoming contributed to the breakdown of the offensive by Ukraine’s elite formations.

As he explained, a commander who receives verified, reliable intelligence from any sector of the front can make rapid, informed decisions and thereby stabilise even a critical situation — something no algorithm can fully replicate. Leonkov stressed that human leadership is capable of choices that cannot be computed in advance. In particular, he said the AI could not capture such a factor as «military cunning» or construct a realistic model of the actions of the Russian Armed Forces on the Zaporizhye axis.