"Soldiers" of the AI generation have been used to spread anti-Ukrainian messages in Europe

EDMO has identified Facebook, Instagram, TikTok, and YouTube accounts that used characters presented as Polish soldiers, but generated with artificial intelligence, for messages about Ukraine, refugees, the EU, Germany, and politics in Poland. The report describes the phenomenon as "synthetic authority," in which AI is used to fabricate individuals who appear to speak from a position of real experience or credibility.

AI-generated characters presented as Polish soldiers have been used on multiple social platforms to distribute messages about Ukraine, refugees, the European Union, Germany, and politics in Poland, according to the European Digital Media Observatory, EDMO. Accounts with names like "Voice of Polish Soldiers," "Voice of the Military," and "Polish Defence" have been identified on Facebook, Instagram, TikTok, and YouTube, and the report describes the technique as the emergence of a form of "synthetic authority."

In short

EDMO has identified accounts that presented AI-generated characters as real Polish soldiers.

The accounts "Voice of Polish Soldiers," "Voice of the Military," and "Polish Defence" were active on Facebook, Instagram, TikTok, and YouTube.

The characters were used for messages about Ukraine, refugees, the EU, Germany, and Polish domestic politics.

EDMO refers to this technique as "synthetic authority," as AI allows for the fabrication of speakers, witnesses, or social roles that seem to have authentic authority.

The report separately identifies the use of an AI-generated voice over authentic images of German politician Alice Weidel to fabricate statements against Ukrainian President Volodymyr Zelensky.

The tactic identified by EDMO is not limited to producing a fake photo or video. Artificial intelligence is used to construct a person who seems to exist and is presented to the public as a representative of a category with authority over the subject discussed.

In Poland, Demagog has identified accounts named "Voice of Polish Soldiers," "Voice of the Military," and "Polish Defence," distributed across four of the largest social platforms: Facebook, Instagram, TikTok, and YouTube.

The characters presented to the public as soldiers were generated with AI and were used to talk about Ukraine and Poland's relationship with the war, as well as about refugees, the European Union, Germany, and themes from Polish domestic politics.

The military appearance is important for the mechanism described in the report. A message about security or war may seem more credible if the person delivering it is perceived as a soldier and, therefore, as having direct experience or access to insider information.

EDMO includes such characters in a broader category of "synthetic and deceptive personas." This includes AI-generated soldiers, false identities of experts, and communities that imitate official groups to fabricate social credibility.

The report describes this evolution as the expansion of "synthetic authority." Artificial intelligence can be used to create speakers, witnesses, and social roles that do not exist but can be presented to the public as legitimate sources.

The technique differs from a simple anonymous account. Instead of hiding the author's identity, the operator deliberately constructs a visual and social identity that suggests experience, profession, and access to relevant information.

Another example identified by EDMO shows that the same technique can be used without the complete creation of a new character. In Georgia, Myth Detector documented a video in Russian that used authentic images of Alice Weidel, co-chair of the German AfD party, over which an AI-generated voice was superimposed.

The fabricated voice attributed to Weidel attacks on Ukrainian President Volodymyr Zelensky that she had not expressed in the authentic material. In this case, the authority of a real person was exploited by artificially modifying what she seemed to say.

The two methods pursue the same result: creating the impression that the message comes from a person who has reasons to be believed. In one case, an entire identity of a soldier is fabricated, while in the other, the voice of a real public figure is fabricated.

EDMO considers the expansion of synthetic authority one of the important findings of July. Generative technology allows for the fabrication not only of content but also of the source that apparently communicates it.

This difference can change the way misinformation is received. The public no longer receives just an uncredited statement, but may see a supposed soldier, expert, or politician who gives it an appearance of personal experience and legitimacy.

In the case of the accounts identified in Poland, distribution across multiple platforms allows the same fabricated identity to appear in different digital ecosystems. Facebook, Instagram, TikTok, and YouTube offer different content formats, but the character can be maintained as a common element of the message.

EDMO includes "synthetic authority" among the main manipulation techniques observed in July 2026. The report shows that artificial intelligence is used to fabricate identities that imitate professional or social categories in which the public may have trust.

Accounts with AI-generated Polish soldiers represent one of the documented examples. They have distributed messages about Ukraine and other political themes on Facebook, Instagram, TikTok, and YouTube.

The phenomenon is distinct from the simple creation of false images: in these cases, AI is used to construct the apparent author of the message and the credibility associated with it.

https://2eu.brussels/ro/news/soldati-generati-cu-ai-au-fost-folositi-pentru-raspandirea-unor-mesaje-anti-ucrainene-in-europa