Purportedly AI-Generated Airport Videos Reportedly Showed Diaspora Voters Opposing Albin Kurti

The Facebook page "Tana t'zezat" allegedly posted three AI-generated videos purporting to show expatriates at Pristina Airport saying they would not vote for Albin Kurti. The videos reportedly mimicked street-interview footage by using microphones styled to resemble Kosovo television broadcasters Klan Kosova and T7, and were reportedly assessed as highly likely to be AI-generated.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.1 Disinformation, surveillance, and influence at scale

The incident involves a coordinated campaign using AI-generated deepfakes and manipulated images to conduct large-scale political disinformation during an election.

Additional risk subdomains

• 3.1 False or misleading information: The AI tools were used to generate fabricated images and videos containing false information about political figures to mislead voters.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The incident was caused by human actors intentionally using AI generation tools post-deployment to create deceptive political content to influence an election.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The report describes AI-generated content such as deepfakes used to manipulate public opinion, which falls under the category of systems requiring transparency obligations so users know they are interacting with AI-generated content.

AI system and alleged parties

• AI system: unspecified
• AI purpose: Deepfake Video Generation; Deepfake Image Generation
• Behaviour type: Tool
• Alleged developer: Synthetic audio technology developers, Deepfake technology developers
• Alleged deployer: Information manipulation actors, Disinformation actors, Deepfake creators
• Alleged harmed parties: Voters in Kosovo, Voters, Journalism, General public of Kosovo, General public, Epistemic integrity, Electoral integrity, Democratic integrity, Albin Kurti

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Minor, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Minor, indirect Negligible
• Democracy: direct Minor, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Minor, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report explicitly describes malicious content created and spread directly by the incident in the form of political deepfakes.

Directly caused: AI-generated deepfake videos and manipulated images targeting political figures were directly created and shared on Facebook to deceive the public.

Inferred additional harm: It is likely that additional unverified AI-generated political content was created and circulated across other social media platforms during the election period.

Civil rights

Reported: The report explicitly describes attempts to manipulate the democratic voting rights of the Kosovo electorate through deceptive AI content.

Directly caused: The electorate's right to free and fair elections was directly targeted by deepfakes designed to mislead voters.

Democracy

Reported: The report explicitly describes the erosion of democratic norms through the use of AI-generated content to manipulate public opinion and influence parliamentary elections.

Directly caused: Deepfake videos and manipulated images of political figures were used to distort political narratives and influence voter perception during the election period.

Inferred additional harm: The widespread use of AI deepfakes in elections likely erodes public trust in democratic institutions and the integrity of electoral processes.

Epistemic

Reported: The report explicitly describes epistemic harm through the AI-driven fabrication of interviews, photos, and statements to erode shared truth during an election.

Directly caused: AI-generated deepfake videos fabricated interviews of diaspora voters, and AI-generated images fabricated meetings between political figures.

Inferred additional harm: The proliferation of such deepfakes makes it harder for citizens to distinguish real news from fabricated content, degrading the overall information ecosystem.

People affected

• Occurrences reported: 1
• People reportedly harmed: 3
• People reportedly exposed: 40000

Potential causes

Management

• Inadequate Risk Assessment: Social networks failed to prepare for AI-driven election interference.
• Poor Integration of AI Detection: Platform managers did not integrate real-time AI detection tools like Hive.

Technology

• Sophisticated AI Video Generation: AI tools allowed creation of highly realistic fake interviews with lip-sync.
• AI Voice Cloning Tools: AI cloned voices to mimic real speech, though they remained slightly robotic.
• AI Image Synthesis Software: Image generation tools created fake photos of politicians and weapon factories.

Data Inputs

• Use of Real Media Logos: Training or blending real logos to deceive viewers into trusting fake videos.
• Source Image Manipulation: Real photos of politicians were input into AI tools to create fake meetings.

Human Factors

• Cognitive Bias of Audiences: Viewers easily believe visual media due to confirmation bias during elections.
• Intentional Malicious Creation: Bad actors deliberately used AI to generate political disinformation.

Process and Methods

• Lack of Proactive Moderation: Facebook failed to proactively detect and block AI-generated political fakes.
• Post-Facto Fact Checking: Fact-checking organizations only verified content after it had already spread.

Regulatory Environment

• Absence of AI Labeling Laws: Lack of regulations requiring clear watermarking or labeling of AI content.
• Weak Platform Liability: No strong legal penalties for social media hosts distributing election fakes.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear, fact-checked details about the specific Facebook pages, the nature of the AI-generated videos and images, and the verification methods used to confirm they were AI-generated. The context of the Kosovo parliamentary elections is well-defined.
• Ambiguities identified: The specific AI models or platforms used to generate the deepfakes and manipulated images are not identified.
• Alternative interpretations: None. The intent to influence the election through fabricated AI content is clear and verified by fact-checkers.

During the December 2025 Kosovo elections, a coordinated campaign used AI deepfakes to target Prime Minister Albin Kurti and manipulate voters. While no physical harm occurred, the incident represents a substantial threat to information security and democratic sovereignty, demonstrating how accessible AI tools can be leveraged to distort electoral processes.

• Overall national security impact: Substantial
• Scope: Single nation
• Primary target: Kosovo
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Near-term. The threat was active during a critical election period, requiring rapid response from fact-checkers to mitigate ongoing impact.
• Autonomy: Human-controlled. AI systems were used strictly as tools by human actors to generate specific deceptive content with no autonomous decision-making.
• Novelty: Evolved capability. While political deepfakes are established, the sophistication of targeting diaspora voters and forging local news logos represents an evolved threat.

Impact by dimension

• Physical security: Negligible. No kinetic threats, physical safety hazards, or critical infrastructure compromise were reported in this incident.
• Information security: Substantial. Coordinated disinformation campaign utilizing sophisticated AI deepfakes and fabricated media logos to target political figures and manipulate voter perception.
• Sovereignty: Substantial. Direct attempt to manipulate democratic parliamentary elections and influence the selection of state leadership through deceptive synthetic media.
• Economic security: Negligible. No impact on financial systems, strategic industries, or technological competitive advantage was reported.
• Societal stability: Minor. Erosion of the information ecosystem and public trust, but did not result in civil unrest or systemic human rights violations.