Nordic Review Warns Election Misinformation Grew More Local, More AI-Powered
In the immediate aftermath of a hard-fought and bitterly contested election season, the Nordic region’s leading disinformation researchers have delivered a stark assessment: the information environment that surrounded voters was poisoned by a mixture of familiar falsehoods and new technological threats. The review, published by NORDIS—one of fifteen EDMO Hubs operating under the European Digital Media Observatory—was produced together with the Swedish fact-checking organisation Källkritikbyrån and the Fojo Media Institute, and it paints a picture of an election marked by coordinated rumour, partisan attacks on electoral integrity, and a wave of synthetic media that blurred the line between reality and fiction. According to the review, the most serious harm did not necessarily come from crude foreign interference but from a more localised, continuous and often automated circulation of misinformation. Social media platforms, encrypted messaging apps and livestreaming services allowed false claims to travel across borders in minutes, often in the native languages of the Nordic countries and tailored to local political sensitivities. The report stresses that while no Nordic election administration reported a successful cyberattack on the vote count itself, the damage was done earlier: millions of voters encountered claims that the election could be rigged, that ballots could be deleted, or that AI-generated videos proved corruption among political elites. The result, the review argues, is an increasingly fragile trust in the very institutions that guarantee free and fair elections.
The first and most familiar theme explored in the NORDIS review is election interference. But unlike the stereotype of a shadowy foreign actor hacking into voting machines, the reality was more diffuse and harder to counter. The review documents how claims about interference came in several overlapping layers. In some cases, state-aligned media and social media accounts amplified allegations that a foreign power—often described as a hostile neighbour—was preparing to manipulate the result. In other cases, domestic political actors used the spectre of foreign interference as a way to cast doubt on the electoral process itself. The review also highlights a particularly dangerous trend: the reinterpretation of legitimate cybersecurity threats as proof of voter manipulation. When an electoral authority in one Nordic country detected a minor online vulnerability, false narratives transformed that routine technical finding into blockbuster allegations of hacking and vote flipping. Such stories were shared widely by partisan news sites, populist politicians and anonymous accounts that often presented themselves as independent commentators. The review notes that these narratives rarely originated from a single source; instead, they were assembled from fragments of true events, manipulated images and out-of-context videos. This made the work of fact-checkers extremely difficult, because each piece of the story had a grain of truth that could be exploited. In one particularly insidious pattern, claims originally made in one Nordic language were translated into another, repackaged by local activists and shared in private groups, making them nearly impossible to trace back to their origin. The review concludes that election interference is no longer only about foreign meddling; it is also about the deliberate erosion of public confidence from within, using the same playbooks that have been developed in other democracies.
The second major category of misinformation identified in the review concerns voter fraud. NORDIS and its partners found that false allegations of widespread voter fraud were among the most viral and emotionally charged pieces of election-related content. Even though the Nordic countries have some of the most transparent and well-administered electoral systems in the world, the review found persistent claims that the vote was not secure. Postal voting was a favourite target, with false stories claiming that mail-in ballots could be forged, lost, or altered after being sent through the post. Other popular falsehoods described deceased people on the electoral rolls, vote-buying operations organised by criminal gangs, and polling stations where election officials were supposedly pressured to manipulate the count. The review is careful to emphasise that these claims were not supported by any evidence uncovered by election observers, national authorities or journalists. Yet the allegations did not need to be true in order to be effective. A single viral post showing a photograph of an unmarked van outside a polling station was enough to spark thousands of shares, questioning the integrity of the entire election. The report’s authors argue that such myths exploit a basic psychological vulnerability: when citizens do not have the time or resources to verify every piece of information, they rely on shortcuts such as partisan identity, social media trust and personal networks. The result, according to the review, is a fragmented information environment in which fewer voters share a common set of facts. The review also warns that voter-fraud narratives can have a chilling effect on democratic participation. If a voter comes to believe that the election is rigged, they may be less likely to vote, or more likely to refuse to accept the result if their preferred candidate loses. In this way, misinformation about voter fraud does more than mislead; it undermines the entire democratic process.
The third and perhaps most novel theme of the NORDIS review is the rapid proliferation of AI-generated content. The review describes 2024 as the first truly generative-AI election, and it offers several examples of how the technology was used to mislead the public. In one common tactic, political candidates were shown in realistic but completely fake videos making incendiary statements. In another, AI-generated audio was used to create phone calls that appeared to be from a party leader, urging voters not to go to the polls. The review also found evidence of AI-generated images depicting election officials in compromising situations, as well as fake social media profiles designed to look like local activists. These synthetic creations were often shared within closed messaging groups, where fact-checkers and journalists had less visibility. By the time they were debunked, they had already been seen by thousands of voters. The review notes that generative AI lowers the cost of producing misinformation to near zero. In previous elections, producing a fake video required sophisticated editing skills and significant computational resources; today, anyone with access to a free online tool can create a convincing deepfake in minutes. This has profound implications for democracy. If every video can be faked, then citizens may be unable or unwilling to believe what they see. Rather than increasing scepticism, the review warns, this environment makes it easier for political actors to dismiss genuine scandals as AI-generated forgeries, a phenomenon sometimes known as the “liar’s dividend.” The review therefore calls for urgent action from technology companies, governments and media regulators to require watermarking, make AI models more transparent and provide reliable tools for identifying synthetic content.
At the heart of the review is a detailed account of how NORDIS, Källkritikbyrån and Fojo Media Institute attempted to confront these challenges during the election. NORDIS, as one of fifteen EDMO Hubs, coordinates research, fact-checking and media-literacy activities across the Nordic countries. Källkritikbyrån, a Swedish organisation that specialises in source criticism and critical thinking, brought its long experience in debunking viral rumours and teaching the public how to verify information. Fojo Media Institute, based at Linnaeus University in Kalmar, contributed its expertise in training journalists and media educators throughout Europe and the world. According to the review, the three organisations operated as an early-warning network, monitoring social media platforms in real time and identifying false stories as they began to trend. They published fact-checks, collaborated with mainstream news organisations, and held live briefings for election officials and civil-society groups. Perhaps most importantly, they developed a cross-border alert system: when a false claim was identified in one Nordic country, the other countries were immediately notified. This was particularly valuable because many of the most successful disinformation campaigns used local language versions of the same narrative, adapting the story to each national context. The review also highlights the importance of media literacy education. In the weeks leading up to the election, NORDIS and its partners ran training sessions with high-school teachers, librarians and community leaders, helping them explain to young people how to assess sources, identify manipulated content and understand the difference between opinion and fact. These efforts were not always easy. The review acknowledges that fact-checking and media-literacy campaigns often struggle to compete with the scale and emotional power of online misinformation. Yet it argues that they are essential counterweights, and that a public equipped with critical-thinking skills is far better prepared to resist manipulation.
The final section of the review looks ahead to future elections and draws a set of difficult lessons. The first lesson is that misinformation is a structural problem, not simply a public-relations problem. It cannot be solved by fact-checking alone; it requires sustained investment in education, regulation and platform accountability. The review urges social media companies to do more to surface reliable information, to label AI-generated content, and to act quickly against accounts that systematically spread falsehoods. It also calls on governments to ensure that online political advertising is transparent and to treat misinformation as a serious threat to democratic institutions. But the review is careful not to place all responsibility on external actors. It insists that journalists, political parties and ordinary citizens also have a role to play. Political leaders should refuse to amplify unverified accusations, and the media should avoid lending credibility to false frames in the name of balance. Voters, meanwhile, must be supported in their efforts to become more resilient to manipulation. The review ends with a warning and a hope: the next election will bring more sophisticated deepfakes, more personalised propaganda and more angry, polarised debate. But the Nordic countries are not starting from zero. Their high levels of public trust, strong institutions and well-organised civil society give them an advantage in the fight against disinformation. The question is whether that trust can be maintained as the pace of technological change accelerates. NORDIS, Källkritikbyrån and Fojo Media Institute make clear that their work is far from over. This review is not an endpoint but a starting point for a broader reflection on how democracies must adapt to a world where information—and misinformation—can be manufactured at unprecedented speed and scale.