Russia serially uses AI drones, Ukraine lags in scaling - "Flash"

Russia is already serially using strike drones with artificial intelligence that can autonomously navigate and recognize objects. Ukraine has its own similar developments and prototypes, but has not yet managed to scale them to mass production.

This was stated by Serhiy Beskrestnov, advisor to the President of Ukraine on technological defense areas, known as Flash, in an interview with DOU.

According to him, the role of artificial intelligence in drone control will only grow. Russian troops are already using V2U UAVs and new versions of "Molniya" that can operate autonomously.

Flash asserts that Russia is developing not only autonomous flight, but also automatic target acquisition. Neural networks on board such drones are trained to recognize various types of objects, including vehicles, railway infrastructure and warehouses.

In this case, the drone can navigate by the camera image and independently identify the specified category of objects even with problems with satellite navigation.

Solutions have already been scaled on the front

Beskrestnov calls the key advantage of Russia not the very existence of such technologies, but their serial application.

According to him, the Russians have installed artificial intelligence systems even on relatively cheap and mass-produced "Molniya" UAVs, and the technology is already being used at scale on the front.

He adds that in Russia this technology has already become a serial solution and is used on a frontline scale.

"Flash" believes that at the current pace, Russia is potentially capable of extending such solutions to thousands or even tens of thousands of drones.

Flash also reported that Russian developers have been building data sets for training neural networks for several years. In his assessment, the next step may be the introduction of similar technologies into long-range strike drones like Shahed.

Ukraine has prototypes, but lacks the necessary scale

Ukrainian developers are also working with neural networks and computer vision systems. According to Beskrestnov, Ukraine began using AI to analyze images from reconnaissance drones and search for camouflaged targets even earlier than Russia.

However, the Russian side, in his assessment, has overtaken Ukraine precisely in moving neural networks directly onto the strike UAV's onboard system and subsequent scaling of the technology.

Ukraine already has prototypes that demonstrate the viability of this approach, but lacks a platform that could be mass-produced in the tens of thousands of units.

Flash considers autonomous systems one of the key future directions of unmanned warfare. The proliferation of electronic warfare and suppression of satellite navigation will force both sides to increasingly move to visual navigation and autonomous object recognition.