How Ukraine Uses AI to Fight Russian Information Operations


Russia’s brutal war in Ukraine is increasingly acknowledged to be the first artificial intelligence war. To be sure, emerging technologies have played an important role in this war, and the information front is certainly no exception. Yet, our understanding of artificial intelligence in information operations remains limited because scholarship often narrowly focuses on AI’s role in corrupting the information space as opposed to defending the information space from corruption. This commentary will therefore highlight what we can learn from Ukraine about how to utilise AI in the fight back against an aggressor.

Take for example the Ukraine-based startups Osavul and Mantis Analytics. Both were born out of the first days of the full-scale invasion and are in the business of detection and counteraction of disinformation campaigns against their country. They are accomplishing this through large language models (LLMs) and natural language processing (NLP) models that scour and analyse the vastness of the Internet for narrative patterns, dramatically speeding up their detection and allowing authorities to get the factual message ahead. This technology tackles the most important variable in how effective an information attack will be – time. With speed in Ukraine’s favour, truth can reach the audience before the Russian narrative does, limiting its spread. Both companies now are working with government agencies including the National Security and Defense Council of Ukraine.

Image: MantisAnalytics program interface demonstrates narratives it identified and their timeline.

Osavul’s strategy addresses the other side of the narrative too; their software can detect coordinated inauthentic behaviour, giving insight to who (or which trolls) the actor might be. These initiatives are born out of Ukraine’s rich technological innovation scene that has now been reoriented and adapted to wartime needs.  

While the AI buzz tends to surround Kyiv’s tech companies and startups, the Ukrainians who might be taking the most advantage of it are journalists. Independent media and digital activists have leveraged artificial intelligence to assist with Open Source Intelligence (OSINT) investigations, investigations that have been central not only to Ukraine’s success in documenting alleged war crimes for legal accountability and historical memory but also in debunking Russian claims about how events have unfolded.

OSINT’s double-edged sword is the sheer amount of data and information journalists have to analyse – a problem perfectly fitted for AI to solve. Whether it’s using satellite imagery as evidence of war crimes, scraping social media images to geolocate combatants, or employing digital verification techniques to authenticate visual evidence, AI supported OSINT counters misinformation related to Russian dilution of facts.

As put by the Royal United Services Institute, “scrutinising each and every event is important for building overarching narratives about the success and legitimacy of the war, especially as Russia seeks to fabricate news”. For this war in particular, OSINT has been crucial for perception management. It’s now an integral part of leading Ukrainian newsrooms like the Kyiv Independent. Examples of powerful investigations conducted with the help of open-source tools can be accessed here and here.

Ukraine’s use of artificial intelligence against Russia’s information war is not without its controversies, however. Although Clearview AI – the facial recognition software engulfed by claims of privacy violations and banned by several governments – is most associated as an aid on the battlefield and in the courtroom, it has also been applied in Ukraine’s counteroffensive against Russian propaganda. In response to the common Kremlin narrative that their side is facing minimal to no losses, the Ukrainian government set up (“Потерь.НЕТ” meaning “no losses” in Russian), a site displaying photos and information on dead Russian soldiers identified by Clearview’s software. The intended audience is the soldiers’ families, with the purpose of awakening Russian society to the true cost of their war. Despite its pro-Ukraine aim, the Clearview AI partnership is strongly opposed by digital rights activists and its use as an offensive information war strategy is even being debated as violation of the Geneva Convention.

Employing facial recognition software in this way is a more gruesome part of Ukraine’s wider humiliation offensive – the trolling campaigns which Ukraine does so infamously well. Whether it’s effective in Russia or actually promotes further motivation to destroy Ukrainians is unclear, but within Ukraine, and campaigns like it undeniably contribute positively to Ukrainian morale, at the very least countering psychological operations.

Other AI-related criticisms directed at Ukraine include the use of AI-generated images by the Ukrainian Parliament’s X (Twitter) account accompanying a message about a Russian missile strike on Dnipro that killed 44 people and injured 79 others. This undermined Ukraine’s credibility on two counts: first, users were sharing the post without realising that the image was not actually from the scene; and second, because of the action’s potential to serve as fuel for Russian propaganda.  

The cases listed above are, of course, not exhaustive examples of how Ukraine integrates AI into the architecture of its information war strategy. Yet they teach globally applicable lessons and demonstrate Ukrainian innovation.

Inside and outside the borders of Ukraine, Russian psyops present a sophisticated and tailored campaign meant to nudge audience perception toward the Russian side. Artificial intelligence is both the perfect tool and weapon on the information front because it mimics the speed with which a decision is made. By doing so, Ukraine is using artificial intelligence to intercept perception creation at its most critical point.



Star tech enterprise: Emerging technologies in Russia’s war on Ukraine. ECFR.

ШІ проти російськихІПСО. Український стартап Osavul навчив нейромережі воювати з пропагандою. Якпродати таку технологію.

Ghioni, R., Taddeo, M.& Floridi, L. Open source intelligence and AI: a systematic review of theGELSI literature. AI & Soc (2023).

OSINT in an age of disinformation warfare. (n.d.). Royal United Services Institute.

Єгошина, В., &Толстякова, К. (2022, December 1). Як добровольці масово ховали мирних містян іхто саме з військових РФ вбивав в Ізюмі? Радіо Свобода.

Caught on camera, traced by phone: the Russian military unit that killed dozens in Bucha. (2023, June29). [Video]. The New York Times.

Harwell, D. (2022, March3). The gory online campaign Ukraine hopes will sow anti-Putin dissent probably violates the Geneva Conventions. Washington Post.

Верховна Радапроілюструвала трагедію в Дніпрі зображенням, яке створив штучний інтелект.Чому це шкідливо? NV New Voice

download here