On November 18, the opposition-discrediting Facebook page “Aravin” posted a video of a protest rally. In the video, a protester is seen with protest stickers reading “Reclaim your voice!” attached to her body. Shortly after, the young person removes the stickers from her face and sits down on a nearby toilet.
The video of the protest rally is created using artificial intelligence (AI) and does not reflect reality.
The disseminated video was AI-generated and, therefore, does not depict the protest held on November 18. Upon closer examination of the footage, it becomes evident that the video contains flaws characteristic of visuals created with AI.
When the video frames are slowed down, these flaws become even more noticeable. At the moment when the young person removes the stickers from her face, the stickers practically disappear and blend into the face. Additionally, the faces of the protesters in the background appear unnatural and distorted.
It is also worth noting that, despite the resemblance to the real stickers of the “Reclaim your voice” protest, the text in the AI-generated video is practically unreadable, especially the first word, “Reclaim.” The inscription resembles the phrase only in terms of its general outline and appearance. The text is particularly distorted in the extreme lower part of the frame. Such text distortions are characteristic of visuals generated by artificial intelligence.
The movements of the young person also look unnatural in the following frames, especially when she sits on the toilet. The sequence of frames gives the impression that the toilet is located at some distance from the protester, but her body moves as if the toilet were right next to her. As a result, the figure moves abruptly and unnaturally from one spot to another. In the same frames, it can be seen that the papers and the person’s hands blend into each other, and the text on the protest posters in the background is indistinct and unclear.
According to the evaluation of TrueMedia.org, a tool designed to identify AI-generated content, the video contains numerous signs of manipulation using artificial intelligence.
About the Source:
The video was disseminated by the opposition-discrediting Facebook page “Aravin.” The page was created on November 12, 2023, but currently, it only has posts dating back to April 2024. Almost all posts on the page are short video Reels.
The AI-generated video is overlaid with the captions reading “Shig Evropashi” (“Deep into Europe”). Videos with analogous captions are distributed by another opposition-discrediting page, “Shig Evropashi.”
Both pages frequently share identical content. Additionally, both pages, particularly “Shig Evropashi,” sponsor their posts. In November alone, the page sponsored 62 posts, all of which were aimed at discrediting protest rallies and the opposition.
The article has been written in the framework of Facebook’s fact-checking program. You can read more about the restrictions that Facebook may impose based on this article via this link. You can find information about appealing or editing our assessment via this link.
Read detailed instructions for editing the article.
Read detailed appeal instructions.
