IN THIS EDITION
ON THE RISE
Weekly Watch of Emerging Disinformation Risks
The World Cup of Hypersexualized Fakes?
Disinformation generally follows information, in particular when a crisis or an important event happens. This is especially true for a specific kind of disinformation: AI-generated content, in particular images and videos, not labeled as such. And the ongoing FIFA World Cup is the umpteenth confirmation.
One category of AI-generated content seems to have gone particularly viral over the past few days: images of beautiful hyper-sexualized female supporters of national teams. Different fact-checking organizations across Europe verified many images of girls with large eyes, small noses, pointed chins, full lips, curvy figures and football national team uniforms – Belgium, Germany, Argentina, Switzerland, Mexico, Spain, Brazil, and so on – all having one thing in common: they do not exist. They are the results of prompts to generative AI tools.
The (main) reason why this avalanche of false content is generated is not a big mystery: thanks to the social media platforms’ monetization programs, content creators can make money out of their posts if they go viral and amass views, likes and shares. And, even more, the accounts on social media platforms are often linked to other websites, like OnlyFans for example, where they can further monetize explicit content.
It is not a novelty – nor something peculiar to social media communication – that images of beautiful women catch the attention of those who view them, men in particular. But while in previous media there were – at least most of the time – some effective countermeasures, platforms are seriously failing to implement their policies and prevent the monetization at scale of disinformation, and in particular of AI-generated content that is not labeled as such.
But what are the consequences? Aside from the general issue of blurring the lines between fiction and reality, these images of hyper-sexualized women can have a negative impact on users, in particular young boys and girls. The former risk growing up with unrealistic, shallow expectations about the other sex and, even worse, the latter risk developing anxiety and depression if they do not fit the beauty stereotypes that are promoted by AI-generated images.
Sometimes even darker dynamics are at work, and users are exposed to further risks. As shown by previous investigations, AI-generated images can be spread online not only to monetize traffic but also to attract and lure users into networks that disseminate far-right nationalist political propaganda and xenophobia, mixed with soft-core pornography.
FIFA World Cup-related disinformation encompasses also other categories, like other kinds of non-labeled AI-generated images that are sometimes intended to be humorous and draw clicks and shares, or conspiracy theories about specific players. In this last category, it is worth mentioning the one about Kylian Mbappé, who left Paris Saint-Germain for Real Madrid two years ago. A recent article falsely claims that he left France because he had been sexually harassed by French President Emmanuel Macron. The article is based on no verified information. It was published on a website mimicking Eurosport’s and impersonating one of its journalists, and it bears all the hallmarks of a pro-Russian disinformation campaign.
ZOOM-IN
A Closer Look at Cases Detected by the EDMO Network
Monetisation of major sports events through mass creation and distribution of AI-slops, and misleading narratives targeting sport fans to saw confusion around sensitive topics such as health or ethnical discrimination, are defining features of today’s disinformation landscape, as illustrated by the following selected cases.
Meta removes fake hockey fan pages amid the 2026 Stanley Cup playoffs
Contrary to the “benign neglect” attitude generally shown so far by social media platforms towards AI-slops exploiting the World Cup for monetisation purposes, CEDMO reports that Meta did remove 39 Facebook pages after an investigation by AFP revealed they were spreading false and misleading content aimed at hockey fans during the 2026 Stanley Cup playoffs. The pages, managed from Vietnam, exploited interest in the NHL to generate engagement and drive traffic to advertisement-heavy websites. Together, they had amassed approximately 244,600 followers before being taken down for violating Meta’s policies.
A common tactic used by these pages was posting fabricated “breaking news” claims that well-known current and former NHL players had been diagnosed with serious illnesses, particularly stage IV glioblastoma. Players and legends falsely targeted included Wayne Gretzky, Mario Lemieux, Mark Messier, Larry Robinson, and others. In some cases, the claims were obviously false, such as alleging illness for active players still competing or for Stan Mikita, who died in 2018. Larry Robinson publicly denied one such rumour. The pages also circulated unverified stories portraying players as making large charitable donations to cancer patients. AFP found no evidence supporting these claims.
No, the United States did not close its border to travellers from Africa due to the threat of Ebola during the 2026 World Cup.
The Chequeado article debunks a viral claim circulating on TikTok that the United States has closed its borders to travellers from Africa during the 2026 World Cup due to an Ebola threat. In reality, there is no general border closure affecting all African travellers. Instead, U.S. authorities have introduced a temporary and limited health-related measure that applies only to individuals who have recently been in specific countries – namely the Democratic Republic of Congo, Uganda, and South Sudan – within the 21 days prior to travel. These restrictions are based on public health protocols and do not amount to a continental ban.
The article also notes that citizens and permanent residents are not affected in the same way, and that such claims are part of a broader wave of misinformation about World Cup-related policies. It emphasizes that neither the U.S., Mexico, nor Canada have closed their borders to African travellers, and urges caution when interpreting viral content online.
No match assignment for referee Artan in Canada is known
In the US, the strict entry requirements have led to an initial problem for World Cup officials as Somali referee Omar Artan was denied entry in Miami due to alleged security concerns. On social media, the decision by US border officials is being contrasted with an alleged message from Canada inviting him to referee in Vancouver, seemingly in opposition to the US’s entry denial.
The claim appears to stem from a social media post by David Eby, a politician from the Canadian province of British Columbia, who wrote that Artan would be welcome in British Columbia and suggested he could referee in Vancouver. However, the post did not explicitly refer to a World Cup match and could be interpreted as a general expression of support. DPA’s fact-check notes that FIFA stated Artan would be unable to train or officiate at the World Cup after being refused entry to the U.S. Consequently, there is no confirmed World Cup assignment for him in Canada. It further notes that, about a year ago, the US government under President Donald Trump published a list of countries whose citizens are barred from entering the United States. Artan’s home country, Somalia, is on that list. A similar regulation was struck down by the Supreme Court during Trump’s first term.
ELECTION BEAT
Tracking electoral disinformation through EDMO Hubs
Fact-check Malta: Did PN leader Alex Borg win more votes than Robert Abela and all previous PN leaders?
Following Malta’s May 31 election, Nationalist Party (PN) supporters have pointed to leader Alex Borg’s vote tally as evidence of personal success despite the party’s defeat, noting he won more votes than Prime Minister Robert Abela and every previous PN leader. A closer look at the figures shows this claim is technically accurate but incomplete. Borg’s 21,825 votes edged out Abela’s 21,682, but Borg ran in Gozo, Malta’s largest district, which inflates raw totals; by vote share, Abela actually led by two percentage points (39.6% to 37.5%).
The “best-ever PN leader” framing also falls apart under scrutiny: both Simon Busuttil (2017) and Eddie Fenech Adami (2003) posted higher vote shares than Borg. He does surpass Bernard Grech and Lawrence Gonzi’s later results, but trails most Labour leaders of the past 25 years, including Joseph Muscat, who twice topped 53%. The figures are real, but the comparison is inaccurate.
Russia ramped up disinformation before Armenia vote
Armenian Prime Minister Nikol Pashinyan has won a new term, with his Civil Contract party securing around 49.8% of the vote on June 7, comfortably ahead of pro-Russian opposition figure Samvel Karapetyan. The result follows what analysts describe as an unusually aggressive Russian disinformation campaign aimed at discrediting Pashinyan’s push toward closer EU ties. In just one week in May, researchers tracked 31 fake reports accusing Pashinyan of election fraud and warmongering, some mimicking outlets like Politico and Armenpress.
One AI-generated video falsely claimed Pashinyan had an “aggressive cancer,” using a cloned Euronews presenter’s voice. Storm-1516, a group linked by French authorities to Russian intelligence, targeted Armenia more than any other country over the past year. Experts say opposition parties have increasingly begun echoing Russian narratives themselves, blurring the line between domestic and foreign disinformation. Moscow denies interference, instead accusing Europe of “dragging” Armenia toward the EU and NATO.
GLOBAL PULSE
Disinformation narratives shaping the world’s conversations
Russia is flooding Ukraine with fake migrant stories – and it’s working
A surge in claims about a supposed influx of foreign migrants in Ukraine is the result of a coordinated disinformation campaign, largely driven by pro‑Russian networks. These narratives falsely suggest migrants are replacing Ukrainians, despite repeated debunks by officials.
The campaign relies on AI‑edited videos, recycled footage and misleading clips, blending fabricated content with real issues like labour shortages to make the story more convincing. By exploiting fears around identity and the war, the messaging aims to divide society and weaken trust, showing how disinformation can amplify tensions by distorting real challenges.
Data poisoning: the invisible threat targeting large AI models
As AI chatbots take on a larger role in how people access information, concerns about “data poisoning” are becoming more pressing. This refers to efforts to insert misleading or harmful content into the data used to train large AI models, with the aim of skewing their responses or behaviour. Because these systems rely on vast amounts of material scraped from the internet, they are inherently vulnerable: manipulated or low‑quality data can be absorbed during training, potentially influencing how models answer questions without users realizing it.
This makes AI tools an increasingly attractive target for influence operations and disinformation campaigns, since shaping the data behind them can indirectly shape what users see. The difficulty of detecting such manipulation highlights a key challenge: ensuring the integrity of training data as AI becomes a central part of the information ecosystem.
Viral “Sex Slavery” Video Misidentified – Actually an Iraqi Performance
A widely shared video claiming to show women being sold as sexual slaves in Syria is misleading. The footage is not real evidence, but a staged artistic performance by an Iraqi artist.
Originally filmed in 2023, the performance depicts women in chains to draw attention to the suffering of Yazidi victims, not to document current events. The case shows how powerful visuals can be stripped of context and reused, turning a piece of art meant to raise awareness into a tool for spreading false or distorted claims.
No Evidence Unvaccinated Children Are “Healthier,” Viral Claim Misleads
A viral claim that unvaccinated children are “much healthier” than vaccinated ones is unsupported. It relies on a supposed study that cannot be verified and lacks credible scientific backing.
The claim originates from a network of low‑credibility or disinformation‑linked websites, where the same content is repeatedly shared without reliable sources. The cited evidence is missing or misrepresented and does not prove any general health advantage.
This reflects a common pattern in health misinformation: unverified claims presented as science and amplified online, risking confusion and undermining trust in vaccination.
ON A DIFFERENT NOTE
[Google] AI summary does not merely display or link to search results but constitutes distinct content attributable to the search engine operator.
– German court holds Google liable for fake AI answers, by Richard Connor on DW
Paolo Cesarini, Editorial Director
Tommaso Canetta, Editor-in Chief
Editorial Staff include Elena Coden, Paula Gori, Elena Maggi
This edition draws in part on automated translation and reflects information available as of 24 June 2026. Later developments may not be included.
