“Doppelgänger”: Media cloning in the service of propaganda and the danger for the Balkans

Disinformation today no longer comes only as vague statuses on social networks or as anonymous portals with exaggerated headlines. It is taking on a more refined form: it sounds convincing, looks professional, imitates reliable sources and, above all, is spread at a speed that makes verification seem slow. In this new phase, two phenomena enter that are changing the rules of the game: deepfakes and media cloning networks, such as the internationally known campaign called “Doppelgänger”.

At first glance, these may seem like problems for “big countries”: France, Germany, Britain. But this is a dangerous illusion. The Western Balkans, with its history of political polarization, raw conflicts in the collective memory, and heavy reliance on social media as a source of news, often functions as a terrain where narratives are quickly imported, easily translated, and made local within hours. In Kosovo, this sensitivity is even stronger because the topics affecting security, the north, dialogue, and interethnic relations are such that a spark is enough to produce a massive reaction.

Deepfake, in essence, is a deception built on an old human advantage: we tend to believe what we see and hear. When a video “shows” someone saying something, or when an audio “sounds” identical to a public figure, our first instinct is to think we are facing proof. Artificial intelligence exploits this reflex: it produces a voice, a mimicry, a look that comes so close to the real thing that suspicion begins to seem like paranoia. But here is the paradox: deepfakes not only damage the truth, they also damage trust in the truth. When the public starts saying “maybe it’s a deepfake,” then even the real facts lose their weight.

The case of Slovakia in 2023 is a clear example of this. Two days before the election, a fake audio clip was distributed that allegedly featured Michal Šimečka discussing vote rigging. Although it was denied, its impact is not measured solely by the question “did it change the result?”, but by the fact that it was released at the most opportune moment for disinformation: on the eve of the vote, when the public is emotionally aroused and when institutional and media reaction is often limited. The analysis by the HKS Misinformation Review highlights precisely this dimension: the deepfake may be sufficient not because it convinces the majority, but because it clouded the atmosphere at a critical moment.

And this logic is easily transferable to the Balkans: all it takes is one "leaked" audio before a political decision, an incident on the ground, or an electoral process, and the damage is done before verification begins.

If deepfake is the deception of “evidence”, Doppelgänger is the deception of “source”. It does not necessarily require complicated video technology, it requires something else: dressing the lie in the costume of credibility. This is done by cloning the appearance of media or institutions, creating pages that look like the original, using almost identical domains and publishing articles that have a “normal” journalistic format. Then, with networks of accounts on Facebook, X or Telegram, the article is pushed until it seems like the topic of the day. At this point, the average reader does not check the domain; they see the logo, they see the format, they see the tone and they take it for granted.

Le Monde has reported on examples where pages have been cloned that imitate French state institutions, publishing false claims designed to generate anger and distrust. These cases are important not only for what they say, but for the way they carry out the deception: they don't ask you to trust an unknown portal, they ask you to trust "a source you already know."

This is where the geopolitical dimension comes in. According to public assessments, Doppelgänger has been described as a campaign pushing pro-Kremlin narratives and aiming to influence public opinion in Europe. US Cyber ​​Command, in a 2024 publication, describes this operation as active since at least May 2022, using cloned websites, fake articles, amplification networks and, in some cases, AI-generated content, with the aim of manipulating public perceptions.

The narratives mentioned in such summaries are usually similar: delegitimizing Ukraine, intimidating the European public through the “cost” of sanctions, and cultivating distrust of governments and the media. So it’s not just “pro” propaganda; it’s “anti-trust” propaganda, against social cohesion, and against the idea that there is a common factual basis.

Why does this affect us in the Balkans, and in Kosovo in particular?

Because our region is burdened with issues where interpretation often drowns out the facts. A small incident, a piece of graffiti, a clash, a short video without context, can be transformed overnight into “evidence” for a grand ethnic or political narrative. And when the public already lives in an environment where trust is fragile, then the manipulation does not need to be perfect, it just needs to be credible enough to ignite a reaction.

In Kosovo, this becomes particularly dangerous when disinformation touches on topics such as security, the north, interethnic relations, the dialogue with Serbia, the role of the EU/US, or economic crises. These are topics where a news headline often creates reality faster than the facts themselves. A cloned page that distributes a “leaked document” or a deepfake audio where a public figure “admits” something can produce political backlash, social tension, or panic, without any verified elements.

In this sense, the Balkans often become a convenient laboratory: not because it is “less intelligent,” but because it is more emotionally charged and polarized, with weak structures of public trust.

Protection is not just a technical matter, but a collective discipline. For the public, the simplest and most valuable step is even more banal: check the URL, check the source, don't rely on the screenshot.

For journalism, the challenge is more serious: not to allow the verification gap to be filled with suggestive language. Because in the age of deepfakes and cloning, the problem is not just “is it true?”, the problem is that a well-packaged lie can have the effect of truth for as long as it produces harm.

Ultimately, the question is not whether these tools will be used here, they are already in global circulation. The real question is how quickly we will build the social reflex that resists manipulation: a culture of verification, a journalism that clearly separates fact from interpretation, and a public that understands that today, “appearance” is no longer a guarantee of truth.

The Geopost