Incident Overview
Between late 2023 and mid-2025, a coordinated wave of deepfake advertisements emerged across Croatia, Bulgaria, Albania, and Kosovo. These ads exploited synthetic media to impersonate high-profile individuals—including footballers, physicians, journalists, and television hosts—to promote unregistered or fraudulent medical products. The campaigns primarily circulated on Facebook, leveraging manipulated videos and AI-generated audio to create false endorsements. The targets included Croatian footballer Luka Modrić, physicians Alemka Markotić (Croatia) and Spas Spaskov (Bulgaria), Albanian emergency physician Skënder Brataj, journalist Blendi Fevziu, cardiologist Spiro Qirko, and journalist Ilir Topi, among others. The products advertised ranged from investment schemes like Immediate Matrix to unapproved treatments such as Hondrosol and Hyper Caps, often falsely claiming medical efficacy or government approval.
The incidents were documented by regional fact-checkers and the AI Incident Database’s Southeastern Europe research initiative. Despite reports, enforcement actions against the advertisers or platforms remained minimal, allowing the scams to persist across multiple countries. In some cases, victims of the misinformation sought medical advice or emergency services based on the false claims, exacerbating the harm caused by the deepfakes.
The campaigns followed a consistent pattern: AI-generated media was used to fabricate endorsements, often paired with fabricated narratives—such as a false murder claim in the case of Dr. Markotić—to lend credibility to the scams. The deepfakes mimicked official branding, including Croatian government designs, further blurring the line between legitimate and fraudulent content.
Possible Root Cause
The primary root cause of these incidents lies in the misuse of deepfake technology and synthetic media generation tools by malicious actors. Developers of AI systems—specifically those enabling deepfake video and audio synthesis—facilitated the creation of these fraudulent advertisements. The technology was deployed by scammers to impersonate trusted figures, exploiting their reputations to promote dubious products. The AI systems implicated include deepfake generation tools and synthetic audio technologies, which were combined to produce convincing yet entirely fabricated endorsements.
The incidents align with the Domain Taxonomy of AI Risks, particularly under the categories of Misinformation and Malicious Actors & Misuse. The harm was not an unintended consequence but a deliberate outcome of deploying AI tools for fraudulent purposes.
The lack of robust enforcement against the advertisers or the platforms hosting the content contributed to the persistence of these incidents. Regional fact-checkers and databases like the AI Incident Database identified the patterns but faced challenges in holding accountable the entities responsible for creating or distributing the deepfakes. The minimal regulatory or platform-level intervention allowed the scams to operate across borders, targeting vulnerable populations seeking medical or financial solutions.
Takeaways
The incidents underscore the growing threat posed by deepfake technology in financial and medical fraud. The misuse of AI-generated media to impersonate authoritative figures—such as doctors, journalists, and athletes—erodes public trust in both healthcare and media institutions. The campaigns also highlighted the difficulty in distinguishing between authentic and synthetic content, particularly when deepfakes are paired with fabricated narratives or official branding.
The role of social media platforms in disseminating these scams cannot be overlooked. Facebook, in particular, served as a primary vector for the distribution of deepfake advertisements, despite its policies against misleading content. This raises questions about the effectiveness of current content moderation strategies in countering AI-driven fraud, especially when the technology evolves to produce increasingly convincing synthetic media.
For policymakers, clinicians, and technology developers, these incidents serve as a case study in the real-world consequences of unchecked AI misuse. The harm extends beyond financial fraud to include misinformation about medical treatments, which can have direct health implications for vulnerable populations. Addressing this threat requires a multi-faceted approach, including stronger enforcement against fraudulent advertisers, improved AI detection tools, and public awareness campaigns to educate users about the risks of deepfake content. The Southeastern Europe region, in particular, may benefit from regional collaboration to share intelligence, coordinate enforcement actions, and develop standardized responses to AI-driven scams.
As deepfake technology becomes more accessible, the potential for abuse grows, particularly in regions with less stringent regulatory frameworks. The cases involving Luka Modrić, Dr. Markotić, and other public figures demonstrate how easily reputations can be exploited for profit, with long-lasting consequences for both the individuals impersonated and the public they serve.
Sources
- incidentdatabase.aihttps://incidentdatabase.ai/cite/1456/
- incidentdatabase.aihttps://incidentdatabase.ai/cite/1459/
- incidentdatabase.aihttps://incidentdatabase.ai/cite/1461/
- incidentdatabase.aihttps://incidentdatabase.ai/cite/1451/
- incidentdatabase.aihttps://incidentdatabase.ai/cite/1452/
Incident summary from the AI Incident Database, used under CC BY-SA 4.0.