Reported AI-Generated Videos Placed Kosovo Parliamentary Candidates in Purportedly Fabricated Campaign Scenes

Videos reported as AI-fabricated circulated on Facebook during Kosovo's June 2026 parliamentary campaign, placing Albin Kurti, Vjosa Osmani, Lumir Abdixhiku, and other candidates in staged scenes of party camaraderie and personal degradation.

Risk classification

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

The incident involved the coordinated creation and dissemination of AI-generated deepfakes of political candidates during an active election campaign, aiming to influence public opinion and political processes.

Additional risk subdomains

• 3.1 False or misleading information: The AI-generated videos presented false and fabricated scenarios of political figures to mislead voters.

Causal factors

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

The incident was caused by human actors who intentionally generated and distributed deepfake videos of political candidates to mislead the public during an election campaign.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The incident involves AI-generated content, specifically deepfakes, which are subject to transparency obligations under the EU AI Act to ensure users are aware they are interacting with synthetic media.

AI system and alleged parties

• AI system: MetaAI (Meta)
• AI purpose: Deepfake Video Generation; Social Media Content Generation
• Behaviour type: Tool
• Alleged developer: Video generation technology developers, Synthetic media technology developers, Meta, Deepfake technology developers
• Alleged deployer: Synthetic media creators, Information manipulation actors in Kosovo, Information manipulation actors, #DIASPORAmeKRYEMINISTRIN, Deepfake creators
• Alleged harmed parties: Voters in Kosovo, Voters, Vjosa Osmani, Lumir Abdixhiku, Kosovo parliamentary candidates, General public of Kosovo, General public, Epistemic integrity, Democratic integrity, Albin Kurti

Harm severity

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 Negligible, 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 describes the creation and spread of fabricated deepfake videos on Facebook targeting political candidates.

Directly caused: AI-generated deepfake videos depicting political candidates in fabricated scenarios were spread on Facebook.

Indirectly caused: N/A

Inferred additional harm: It is likely that additional political deepfakes were generated and circulated during the campaign period that were not captured in this specific report.

Democracy

Reported: The report describes the circulation of fabricated videos of political candidates during an active election campaign, which threatens democratic processes.

Directly caused: The integrity of the election campaign was challenged by the dissemination of AI-generated disinformation targeting key political figures.

Indirectly caused: N/A

Inferred additional harm: Such campaigns risk undermining public trust in democratic processes, political institutions, and election outcomes.

Epistemic

Reported: The report explicitly describes epistemic harm through the fabrication of videos showing political candidates in false scenarios.

Directly caused: Multiple deepfake videos were created and shared on Facebook, presenting false narratives about political candidates.

Indirectly caused: N/A

Inferred additional harm: The spread of these videos likely misled some voters and contributed to the erosion of shared truth during the election period.

People affected

• Occurrences reported: 3
• People reportedly harmed: 3

Potential causes

Management

• Growth Prioritized Over Safety: AI companies release tools without sufficient safeguards against political abuse.

Technology

• Easy Generative AI Access: MetaAI and other tools allow users to easily generate realistic fake media.
• Inadequate Platform Guardrails: Generative tools lack strict filters to block generating political figures.

Data Inputs

• Abundant Public Political Media: Large volumes of public images of politicians are available to train AI.

Human Factors

• Intent to Mislead Voters: Actors maliciously generate fakes to influence public opinion in elections.
• Low User Digital Literacy: Social media users share fabricated content without verifying its origin.

Process and Methods

• Reactive Content Moderation: Platforms rely on post-upload reporting instead of proactive AI detection.
• Delayed Fact-Checking Process: Verification organizations only flag fakes after they have spread widely.

Regulatory Environment

• Weak Deepfake Regulations: Lack of legal frameworks to penalize AI-generated political disinformation.

Information quality

• Classification confidence: High
• Reason for confidence: The report clearly details the nature of the deepfakes, the political figures targeted, the platform used (MetaAI), and the fact-checking process that exposed the fabrication. There is little ambiguity regarding the events described.
• Ambiguities identified: The exact scale of dissemination (view counts, shares) is not fully detailed in the report.
• Alternative interpretations: None.

During the June 2026 election campaign in Kosovo, AI-generated deepfakes of political candidates in mundane scenarios were circulated on Facebook. While targeting key political figures during an election, the benign nature of the content and quick debunking by fact-checkers limited the national security impact to a minor level.

• Overall national security impact: Minor
• Response level: Moderate
• Scope: Single nation
• Primary target: Kosovo
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. Represents an ongoing strategic concern regarding the use of synthetic media in democratic elections.
• Autonomy: Human-controlled. The AI system acted strictly as a tool controlled by human actors to generate specific media.
• Novelty: Established threat. The use of AI-generated deepfakes to target political figures during elections is an established threat methodology.

Impact by dimension

• Physical security: Negligible. The incident involved social media deepfakes with no kinetic, physical, or critical infrastructure impact.
• Information security: Minor. AI-generated deepfakes targeted political candidates during an election campaign, representing a minor information operations threat with relatively low-gravity content.
• Sovereignty: Minor. The deepfakes targeted candidates during a parliamentary election campaign, representing a minor attempt to manipulate democratic processes.
• Economic security: Negligible. No economic or technological security assets were targeted or affected by this incident.
• Societal stability: Minor. The propagation of synthetic media during an election risks misleading voters, but the mundane nature of the content limits its threat to societal stability.