Demoralizing the enemy, its military and civilian population, is an important component of military strategy. Russians are interested in ensuring that as many Ukrainians as possible in the rear and at the front are gripped by fear, despair, apathy, and unwillingness to fight for their own state. The Center for Strategic Communications records the coordinated spread of videos generated using artificial intelligence. Their task is to demoralize Ukrainian citizens.
What is happening?
The Ukrainian segment of social networks is being filled with short video appeals from “typical Ukrainians”: military personnel, women, men of draft age with emotional stories about the meaninglessness of defending Ukraine and armed resistance to Russian aggression.
Video appeals from virtual “draft evaders” and their “relatives”
Among the video characters, recognizable types can be identified:
tired, disillusioned soldier;
young man of draft age who illegally crossed the border or plans to do so;
mother who is afraid of losing her son and therefore sent him abroad;
wife or girlfriend whose husband or boyfriend is at the front or fled from mobilization abroad;
corrupt official, head of TCC (Territorial Recruitment Center).
This is how AI depicts corrupt officials and TCC heads
The manipulative statements of the characters are meant to convince the audience that:
Ukraine is a “bad state” that has “given nothing” to its citizens;
Ukrainian military at the front are “defending corrupt officials,” not their own state;
unauthorized abandonment of unit (AWOL), desertion, illegal border crossing are “normal” because “everyone does it.”
How does it work?
The videos are published on Facebook pages that were activated in November-December 2025. Three identified pages with AI videos were created earlier under different names.ні 2025 року. Три виявлених сторінки з ШІ-відео було створено раніше під іншими назвами.
The pages have no remnants of publications before renaming; previous content was likely deleted. One page retained an artifact from its “previous life” – an advertisement for an online betting site in Turkish. Users were offered to bet on the result of the Italy-Turkey national football team match at Euro-2020. The match took place on June 11, 2021, the page was created on June 9 – two days before the game.
Fact-checkers from the “NotaYenota” project found that the LidiiaKhudoshina account (the previous name of the page with AI videos) is active on Facebook but belongs to a bot. The profile photo was stolen from the page of a real person – stylist Yanina Tumarinsson from Kremenchuk.
Content is created according to typical scenarios: mostly selfie videos with appeals to the audience retelling a “life story” that integrates messages about the “bad state,” “corrupt officials” and calls to “get out” from positions (for military) or from the country (for civilians).
Video appeals from “demoralized soldiers”
The videos look quite realistic. Moreover, they are created to inspire audience trust: recognizable types in typical locations directly address viewers or depict “overheard” conversations.
Video appeals from virtual “draft evaders,” “escapees” and their “relatives”
Each video gets tens to hundreds of thousands, sometimes over a million views, a significant number of likes, comments and reposts. Videos are re-uploaded to other platforms: TikTok, Telegram channels, Viber chats, etc.
It’s enough to watch one or several such videos, write a comment, put a like – and social network algorithms will offer more such content as potentially interesting for the user.
How to recognize a fake?
A generated video can be distinguished from a real one using special programs (AI detectors), including online products: both paid and free.
The analysis result of videos from the analyzed Facebook pages was unambiguous: AI detectors estimate the probability of generation at 85-90% and higher.
Generated videos can also be revealed by:
violations of the laws of physics;
unnatural movements of characters;
incorrect stress and unnatural intonations in speech;
presence of disclaimers and watermarks that free versions of programs used to create videos may leave (Sora, Veo and others).
If it’s a Facebook publication, it’s worth paying attention to the page, checking its transparency and other publications.
Analysis of pages with videos revealed several “red flags” at once:
page renaming with “cleaning” of previous publications;
pages are anonymous, as is the video content (characters don’t name themselves, the time and date of events they “tell about” cannot be established);
videos have typical scenarios, all videos are 14-15 seconds long;
there are several videos created, obviously, from the same prompt: “woman shows cellar with preserves where she hid her son,” “sad older woman in an apartment talks about her son’s escape abroad,” “woman on the phone tells a friend about her son’s departure abroad,” “soldier talks about fatigue while sitting,” “soldier walks through the forest and curses the authorities and command,” etc.
What does the enemy seek?
Spreading such videos is aimed at:
normalizing draft evasion, AWOL, illegal border crossing;
society’s perception of these offenses as socially approved behavioral models;
provoking artificial conflicts in Ukrainian society, particularly between military and civilians, military and police, rank and file and command staff of the Armed Forces, front and rear, etc.;
manipulating the emotions of military personnel, those liable for military service and their relatives, stimulating them to commit offenses.
Virtual “soldiers” are “outraged” by the actions of police, TCCs and Ukrainians vacationing in Bukovel
The enemy wants a Ukrainian soldier, after watching such videos, to listen to the advice of a virtual “comrade” and also go AWOL. For mothers and wives to pressure their sons and husbands and encourage them to flee from the military or from the country (because this “help” in AI videos is promoted as a manifestation of “true love”).
The aggressor’s key goal is to deprive Ukraine of the ability to resist through demoralizing Ukrainian society.
What to do?
The Center for Strategic Communications recommends following information security rules:
do not share anonymous videos and posts, even if they emotionally “affected” you;
do not like or comment on suspicious content, do not help its spread;
check page transparency and verify the authenticity of images and videos using online resources for detecting AI-generated content;
report pages with manipulative AI content on social networks.
If you have found a spelling error, please, notify us by selecting that text and pressing Ctrl+Enter.
