Russia is deploying AI-generated "Ukrainian soldiers", "rabbis", and other fake authority figures on Ukrainian social media at scale. A Hromadske investigation, drawing on Ukrainian analytics firm LetsData and Ukrainian intelligence group Molfar Intelligence Institute, has documented three viral AI-generated propaganda videos targeting Ukrainian audiences in early May 2026.
The tactic exploits a structural information-warfare advantage that generative AI has handed to hostile propaganda operations. AI-generated lies spread in seconds, while refutations from human fact-checkers appear hours or days later.
Molfar Intelligence Institute describes a time gap as "structural" rather than incidental. Russia is exploiting it by using what LetsData calls "synthetic authoritative figures", AI-generated characters presented as members of trusted social groups whose fake statements carry perceived legitimacy that anonymous posts don't.
Evolution of lies
The tactic sits within a documented multi-year evolution of Russian AI-disinformation targeting Ukraine. Ukrainian outlet Texty investigated a 2,000-account fake TikTok network running the anti-mobilization hashtag #ценемояукраїна in mid-2024, and AFP documented AI-generated fake surrender videos flooding social media around Pokrovsk in November 2025.
What is new in 2026 is the industrialization of production and the deliberate choice of "authoritative" identities.
Three viral videos, three distinct Russian narratives — all fake
The Facebook page Soldatska Pravda ("Soldier's Truth") posted an AI-generated video on 7 May 2026 that accumulated over 915,000 views.
In the video, a fabricated Ukrainian soldier says he has not been home in two years and asks whether his sacrifice is necessary if "they" — implied Ukrainian politicians — are building "a third house on the French Riviera." The narrative: soldiers sacrifice while politicians enrich themselves abroad.
A TikTok video posted on 1 May 2026 accumulated 557,000 views.
An AI-generated interviewer asks a fabricated rabbi about Ukrainian draft dodgers. Then, the rabbi says those avoiding military service should be stripped of Ukrainian citizenship, and adds that Ukraine belongs to "us, real Ukrainians."
The video operates the antisemitic trope that Ukraine belongs to Jews, a specifically inflammatory frame designed to circulate through both Ukrainian domestic audiences (as apparent evidence of Jewish overreach) and Russian audiences (as apparent evidence of Ukrainian antisemitism). Both readings serve Russian propaganda goals.
A TikTok video posted on 30 April 2026 accumulated 425,000 views.
An AI-generated soldier claims Zelenskyy will not stop the war until "no one is left — neither in the trenches, nor the villages, nor the cities" — and only then will he declare victory. The video transforms societal war-exhaustion messaging into a personalized accusation against the head of state.
Detection markers and media-literacy defense
Molfar Intelligence Institute identifies specific visual and behavioral markers of AI-generated propaganda videos.
- Visual defects: artificial blinking patterns, sound-lip desynchronization, and absence of micro-expressions in the face.
- Source anonymity: characters presented without names, positions, or unit affiliations — what Molfar calls "no-names" — because AI can generate a face but not a verifiable identity.
- Synchronized dissemination: the same video or same-topic videos appearing simultaneously on dozens of new or previously empty social media accounts.
- Artificial promotion (called "warming" in Ukrainian OSINT vocabulary): hundreds of views, likes, and formulaic uniform comments in the first hours after posting, generated by empty accounts, stolen accounts, or foreign-language accounts.
- Cross-platform posting: identical videos uploaded simultaneously to TikTok, YouTube, Facebook, Telegram, and X.
Ukrainian cybersecurity expert Kostiantyn Korsun says the main line of defense remains critical thinking and common sense. Users should verify claims through independent sources. If a video references BBC data, check the BBC directly.
Common sense applies at the basic level: being told you have won an apartment in Kyiv without entering any lottery is a signal to doubt, not to click.
