Deepfakes have become increasingly widespread over the past two years, largely due to rapid advances in artificial intelligence technology. Like many other countries, Kosovo is now facing the growing impact of this phenomenon ahead of the June 7 elections.
The rising number of AI-generated deepfake videos circulating on social media especially on TikTok is raising concerns about manipulation, misinformation, and public trust. According to an analysis published by Sbunker, deepfakes are becoming an increasingly influential tool in political, social, and economic spheres.
A study examining Romania’s 2024 elections, which analyzed more than 400 misleading materials, found that deepfake-driven disinformation campaigns had serious consequences for information integrity and trust in institutions.
While political manipulation during past election campaigns often relied on “cheapfakes” — edited or misleading content created without AI — the emergence of accessible AI tools has made sophisticated deepfakes far easier to produce and distribute. Many citizens in Kosovo continue to engage with and believe such content, highlighting the potential impact on public opinion and democratic processes.
Deepfake Cases Before Kosovo’s June 7 Elections
In recent weeks, multiple deepfake videos portraying politicians making false statements or engaging in fabricated actions have circulated online.
One widely shared example involved Kosovo’s former president Vjosa Osmani and acting Prime Minister Albin Kurti. Published by “Kosova Info,” the manipulated video reportedly reached nearly 750,000 views across different social media platforms.
Another deepfake circulated following reports about cooperation between Përparim Rama and the Democratic Party of Kosovo in Prishtina Municipality, falsely depicting a conversation between Lumir Abdixhiku and Bedri Hamza.
Although some AI-generated audio is easier to detect due to unnatural accents or pronunciation, many manipulated videos continue to gain significant traction online.
“DeepFakeMburoja” Proposed as a Protective Model
To help citizens defend themselves against AI-generated manipulation, experts have promoted a six-step awareness model known as “DeepFakeMburoja” (“DeepFakeShield”).
The model encourages citizens to:
- Avoid rushing to conclusions
- Verify the source of information
- Cross-check content through multiple media outlets
- Watch for technical inconsistencies in videos or audio
- Use digital verification tools such as reverse image searches and metadata analysis
- Think critically about context and motives behind the content
Analysts warn that deepfakes represent not only a technological challenge, but also a major democratic concern, particularly during sensitive electoral periods.
