Throughout August, BIRN monitors recorded cases where AI was misused to impersonate politicians, doctors, journalists, citizens, and public institutions across Southeast Europe.
According to Eurostat, in most countries in the region, less than half of their populations are equipped with basic digital skills, which is well below the EU average. This leaves many people vulnerable to AI-driven deception, and to serious personal and financial risks.
At a global level, OpenAI CEO Sam Altman recently warned of an “impending fraud crisis,” in which nearly anyone could be convincingly imitated, including their voice and appearance, using advanced AI tools like those his company, among others, develops.
These concerns are echoed in many of the cases identified by BIRN monitoring during August.
In Kosovo, AI tools are increasingly misused to manipulate public opinion and discredit political figures. In August, a fabricated video falsely showed Vetëvendosje party member Avni Dehari speaking Serbian in parliament, alongside acting PM Albin Kurti dancing, and a Serbian flag, none of which occurred. Another doctored image depicted Kurti holding a banner reading “Kosovo is Serbia,” with Donald Trump in the background. They reached thousands of viewers, potentially stoking ethnic tensions and undermining political credibility.
In a health-related disinformation case, a fake video falsely claimed that a Kosovar cardiologist had discovered a “home cure” for hypertension. Styled to mimic Albanian public television, RTSH, news broadcasts, it used AI to manipulate footage of an Albanian anchor and a university rector.
In Serbia, an AI-generated video targeted a lawyer, Ivan Ninic, who had offered pro bono legal services to student protesters. The video alleged a connection between Ninic and so-called “Albanian occupiers in Kosovo and Metohija”. Ninic, drawing parallels to the politically motivated assassination of Kosovo Serb politician Oliver Ivanovic, said he feared for his safety. In response, he filed a criminal complaint with the High-Tech Crime Department in Belgrade.
United Group founder Dragan Solak has meanwhile also become the sole target of a malicious AI-driven smear campaign on a newly created X account. The account, which exclusively posts about Solak, spreads fabricated, AI-generated content, accusing him of plotting to shut down United Media’s critical TV channels, Nova S and N1. Fears for media freedom in Serbia are rising following a report about claims that President Aleksandar Vucic tried to force the dismissal of a United Media director.
In Romania, AI-driven disinformation has taken both financial and political turns. In August, the National Cyber Security Directorate issued a warning about a fraudulent investment scam using a deepfake of the Romanian president, Nicușor Dan. The video featured the president endorsing fake government investment schemes, redirecting citizens to fake government websites, where they were asked to submit their personal and financial data.
Another deepfake video of the Romanian president went viral, this time of him mocking citizens living in poverty. The fabricated clip included vulgar and insulting remarks attributed to the president, timed strategically during a politically tense period involving austerity measures. It was widely shared and triggered a public backlash, but was eventually debunked by fact-checkers.
Meanwhile, a separate deepfake video attempted to create panic by falsely showing Prime Minister Ilie Bolojan announcing Romania’s entry into the Ukraine conflict.
In North Macedonia, Prime Minister Hristijan Mickoski’s identity was used in a deepfake promoting a fraudulent investment platform, misusing real interviews, doctored audio, and redirecting users to phishing sites. Similarly, in Montenegro, a deepfake video circulated showing Finance Minister Novica Vukovic allegedly urging viewers to invest in a scheme during a fabricated interview with a journalist.
President Gordana Siljanovska Davkova of North Macedonia also appeared in a deepfake video, misnaming North Macedonia’s most important national holiday, Ilinden, so portraying her as unpatriotic and out of touch.
A more commercial example of misuse occurred when singer Tijana Dapcevic and a doctor, Igor Nikolov, were falsely featured in an ad for urinary incontinence medication. The AI-generated video fabricated testimonials and a fake interview, misleading users into providing their personal and financial information to purchase the medication. Nikolov denied involvement and reported the case.
An AI-doctored video falsely showing Macedonian news anchor Slavica Arsova reporting on a fake money-making video game created by Cristiano Ronaldo circulated on social media, with fact-checkers confirming it as fraudulent.
In Bosnia and Herzegovina, a TV host from Klix.ba warned viewers about a deepfake video that falsely depicted her endorsing a home remedy for vision improvement, using AI-generated visuals and cloned voice.
Throughout the year, the same country also reported cases of AI use in gendered digital abuse, including female minors being blackmailed with “undress” images created using so-called nudify apps.
As documented by BIRN monitoring, AI is increasingly exploited across the region to spread disinformation, commit financial crimes, and discredit public figures. With fraud becoming more sophisticated and harder to detect, and a flooded information space, the lines between what is real and fake are fading. This troubling trend of manipulated realities portends a trust crisis, and underscores the need for regulation, media and digital literacy, and AI transparency across Southeast Europe.