Ukraine Trained AI for Its ‘Spiders Web’ Airfield Drone Attacks at Aviation Museum
Ukraine’s ability to smuggle almost 120 first-person view (FPV) drones into the heart of Russia and wreak havoc on Moscow’s strategic bomber fleet – which has continually launched its missiles against civilian targets in Ukraine, and is also one leg of the Kremlin’s nuclear triad – has undoubtedly made the President Vladimir Putin along with the rest of the world sit up and take notice.
On Sunday, June 1, an audacious long-distance operation mounted by Ukraine’s intelligence services attacked at least five airfields, in Russia’s Murmansk, Irkutsk, Ryazan, and Amur regions, where strategic bombers and surveillance aircraft were located. While the numbers of aircraft destroyed and damaged is still being debated – Kyiv has said it could be as many as 40 – this success has more than bloodied the Russian aerospace forces’ nose.
In July the milblogger “Clash Report” said that Ukraine’s intelligence services were training artificial intelligence (AI) systems that would enable its drones to recognize enemy equipment, including Soviet-era bombers, using images obtained from military museums. Those who asked “so what” at the time got their answer on Sunday when the SBU’s “Operation Pavutyna (Spider’s Web)” decimated a large number of Moscow’s strategic fleet of aircraft deep inside Russian territory.
As Ukrainian President Volodymyr Zelensky made clear, when speaking about the success of the mission on Sunday evening, it had taken 18 months to prepare. The development of the plan reportedly began some time in 2023 when, according to the Ukrainian news site Antikor, operators from Ukraine’s military intelligence directorate (HUR) made hundreds of images of Russian strategic bombers, stored at the Poltava Museum of Heavy Bomber Aviation, from every conceivable angle.
This data was used to identify the most vulnerable areas of the bombers from which to create AI algorithms that would allow weapons, in this case FPV drones, to independently recognize and engage targets. As videos captured during Sunday’s attacks attests the drones didn’t simply crash into their targets but went for the areas where maximum damage was most likely to be caused – weapons pylons carrying cruise missiles and over wing fuel tanks.
How to train AI for military use
There are several stages to training AI to behave as it did during Sunday’s attack – which combined with the secrecy, technical, logistic and operational preparations necessary for it to succeed explains why it took the best part of two years to implement.
The first stage is selection of the right AI algorithm model and architecture to carry out the intended task and to identify the data in the format needed to be provided to the system.
The next stage is data preparation – the gathering of a comprehensive dataset with which to prime the system, cleaning it up and converting it into a format the chosen AI model would recognize.
This is followed by training the AI which involves the repetitive manipulation, feeding and fine tuning of the data and the AI model, which IT nerds call “epochs,” to minimize errors and improve accuracy.
Once training seems to be on track the next stage is validation and testing – seeing how the trained model performs when presented with previously unseen data, such as target aircraft being viewed from various angles, in different lighting and weather conditions. Has the AI truly assimilated the data or has it just learnt limited numbers of recognizable patterns?
In the build-up to operational deployment the system is continually worked on with updates of any new data and adjustments needed to get the maximum performance from the system as possible before it is unleashed for real.
Ukraine has had a reputation as a premium information technology center for decades before Russia’s full- scale invasion – witness the number of gaming companies and notorious hacker groups based in Ukraine or using its experts. The war has brought those skills into sharp focus and the success of Sunday’s attacks on Moscow’s bomber fleet shows it now also a leader in the use of AI for military purposes.
The Poltava Museum of heavy bomber aviation
The museum is sited on an airfield close to the central Ukrainian city of Poltava. It was established in the 1920s where it became Ukraine’s first base for passenger airlines before conversion into a strategic bomber airbase after World War II.
Although it was formally turned into a museum in September 2016 it began to display aircraft in areas close to the training buildings of the 185th Guards Heavy Bomber Regiment in 1987 which slowly began to expand from that date.
It now consists of several open air and exhibition halls that tells the history of heavy bombing, anti-aircraft, artillery and the military units that manned them from the Soviet era until the present day.
Its unique display includes helicopters, transport and fighter aircraft along with a whole range of Tupolev, Sukhoi and Antonov bombers including the Tu-160 “White Swan” (NATO: Blackjack) and the only remaining Tu-95MS (NATO: Bear) aircraft held by Ukraine.
It is perhaps ironic (although probably coincidental) that on June 1, 1996. Ukraine handed over the last of its nuclear warheads and most of its strategic bombers under the auspices of the so-called Budapest Memorandum and, almost exactly 29 years later on June 1, 2025, Kyiv put a significant number of those aircraft “beyond use.”
Source: Steve Brown