Here is a humanized summary of the report in six paragraphs.
—
From 15 to 30 September 2026, a team of researchers and fact-checkers from BROD, the Bulgarian partner in the European Digital Media Observatory network, spent two weeks doing something unusual. Instead of reacting to one fake post at a time, they took a deliberate step back and mapped the entire election information environment as a living, breathing ecosystem. The result is a report built on seven separate datasets, collected exactly one month before Bulgarians headed to the polls for the presidential election on 25 October 2026. If you think of the internet as a crowded town square during a campaign, BROD did not simply stand in the center and listen to the loudest speaker. They walked through every side street, counted the microphones, listened to the whispered gossip, watched which direction people turned when a rumor started, and then compared what they heard online with what was being said on traditional television and radio. The study is organized around a six-layer analytical framework called the IWH methodology, a deliberately structured way of looking at the information landscape from multiple angles. The layers include a quantitative audit of public content across major platforms, a network analysis that tracks how claims move from one account to another, a qualitative reading of the most viral themes and narratives, a deep dive into problematic actors and coordinated behavior, a closer look at mainstream media coverage, and an assessment of which audience groups are most exposed to harm. For an ordinary voter scrolling through social media, all this may sound like academic jargon. But underneath the technical language is a very simple and human question: when Bulgarians step into the voting booth, what are they actually voting on, and how much of what they think they know about the candidates is true?
The six-layer methodology is not just a complicated way to gather data; it is really a way to see the full picture without being fooled by any single source. The seven datasets are like seven windows into the same crowded room. One window shows the stream of news articles produced by Bulgarian media outlets during the monitoring period, including the headlines, the photographs, and the framing choices made by editors. Another window shows the conversation on social networks—the memes, the angry comments, the enthusiastic shares, and the strange, sudden viral moments that seem to come from nowhere. A third window opens onto public Telegram channels, where anonymous users often set the agenda before mainstream outlets even notice. A fourth window focuses on advertising, particularly political and issue-based advertising, because paid promotion can sometimes tell a more honest story about strategy than organic content. A fifth window contains fact-checking databases, which record the claims that were verified or debunked during the campaign, and therefore reveal the most common types of manipulation. The sixth and seventh windows, in turn, look at the structural connections between accounts and the psychological or informational vulnerability of the audiences themselves. This is what makes the report feel so refreshingly human. It does not treat the internet as a faceless machine. It treats it as a place where real people share fears, hopes, and half-remembered news items with their neighbors, their grandparents, and their childhood friends. By using all six layers together, the researchers can see not only what is being said, but also who is saying it, why they might be saying it, and who is most likely to believe it.
When the datasets are combined, a series of recurring narratives emerges from the Bulgarian pre-election period. The most persistent themes are not always directly about candidates. Instead, they often revolve around deeper emotional subjects: national sovereignty, corruption, economic insecurity, demographic decline, and the fear of being ignored by Brussels and the wider European Union. A significant portion of the content frames the election not as a choice between two political personalities, but as a kind of existential battle between the people and a corrupt elite. In this narrative, the specific policies of the candidates matter less than their perceived loyalty to the nation. Another major theme involves the war in Ukraine and Bulgaria’s position in the world, with claims that have circulated for years finding new life in the election context. Some posts describe Bulgaria as a victim of foreign influence, while others warn against the danger of being dragged into a conflict that does not belong to the country. Migration also appears consistently, often with dramatic imagery that is difficult to verify. The emotional core of these narratives is fear: fear of losing national identity, fear of economic collapse, fear of being manipulated by outside forces. What makes the report so valuable is that it does not simply list these narratives as bad or false. Instead, it shows how each narrative fits into a larger emotional landscape. A grandmother in a small town may share a video about rising prices because she is genuinely worried about her pension. A young student may share a meme about national pride because he feels disconnected from the political system. The narratives are powerful precisely because they connect to real anxieties, and that is why they spread so quickly across platforms.
The tactical side of the report reads like a handbook of digital manipulation. Among the most striking findings are the ways in which artificially generated content, particularly images and voice recordings, have become cheap and widely used tools of political communication. The report points to examples of fake photographs designed to make politicians look ridiculous, synthetic audio clips that put false words into candidates’ mouths, and recycled video footage from other countries or from entirely different events presented as if it featured Bulgarian politicians. These forms of content are especially dangerous because they are easy to share and difficult to debunk in the moment. A voter scrolling through a social media feed may not stop to ask whether a picture was created by an algorithm, especially if the picture confirms what the voter already believes. Another worrying tactic is the use of messaging applications, especially Telegram, where anonymous channels can spread information to thousands of subscribers without any meaningful moderation. The report shows how a claim can appear in a small niche Telegram channel, then be picked up by a more active channel, then suddenly appear on Facebook and TikTok, and eventually be quoted by a local news website as if it were a real fact. This kind of cross-platform amplification creates a feeling of inevitability. By the time a fact-checker publishes a debunking article, the story has already traveled through multiple communities and taken root in many people’s minds. The report also identifies coordinated attempts to create a false sense of consensus through the use of repeated comments, identical hashtags, and synchronized posting patterns. These are not just organic conversations; they are carefully orchestrated campaigns.
Behind the tactics, the report paints a nuanced picture of the actors involved. Some are external, notably pro-Kremlin outlets and anonymous operators who have been active in the region for years. Their goal appears to be less about supporting a specific candidate and more about weakening public confidence in democratic institutions. Their content often promotes the idea that the election itself is rigged, that all politicians are corrupt, and that voting is a meaningless gesture. This kind of narrative, if believed widely enough, can discourage participation and create the exact kind of apathy that hostile actors want. At the same time, the report does not ignore domestic political actors. Domestic parties and individuals also use manipulative tactics, sometimes borrowing from the same playbook. The report suggests that the line between foreign and domestic disinformation is increasingly blurred. Local political forces sometimes share content originally produced by external actors, while foreign actors piggyback on real local grievances to make their own narratives more believable. The report also highlights the vulnerability of different audience groups in Bulgaria. Older people, who often rely on television and family messaging groups, are more likely to encounter manipulated content that fits into familiar worldview patterns. Younger voters, by contrast, live on TikTok and Instagram, where short videos and algorithmic repetition can make a false claim feel like common sense. The constant use of irony and humor in online content makes it difficult for many people to tell whether a post is meant to be a joke, a real political statement, or a deliberate piece of manipulation. What is clear is that the threat is not distributed equally. The report’s real contribution here is to remember that behind every metric there is a human being.
Finally, the report ends with a set of recommendations that are meant to protect the integrity of the election without treating voters as fools. It calls for a stronger investment in media literacy, not just in schools, but also in the kind of informal learning that happens in libraries, community centers, and retirement clubs. It encourages fact-checking organizations to publish their findings faster, in simpler language, and in the same formats and platforms where misinformation actually lives. It asks social media companies to be more transparent about the sources of paid political advertising and more responsive to coordinated inauthentic behavior. It also suggests that Bulgarian media outlets, particularly local ones, should avoid amplifying unverified claims simply because they are dramatic or likely to attract clicks. In this sense, the report is not a story of doom. It is a reminder that the election information environment is made by people, and that people also have the power to make it better. The researchers emphasize that while the days before the presidential election on 25 October 2026 are a time of high risk, they are also a time of opportunity for responsible journalism, for open conversation, for humble correction of mistakes, and for citizens to stop and think before sharing. The report does not pretend that the problem of disinformation can be solved overnight. Instead, it gently asks all of us to become a little more curious, a little more patient, and a little more willing to verify what we see online. That is probably the most human response to a complex and sometimes frightening digital world.