Reported AI-Generated Image Portrayed Vetëvendosje Official Dejona Mihali Meeting Aleksandar Vučić

The Facebook page Gjithqka rreth Politikes reportedly published an image falsely suggesting that Dejona Mihali, a senior Vetëvendosje official, coordinated with Serbian President Aleksandar Vučić against Albin Kurti during Kosovo's snap-election campaign. The image was reportedly likely AI-generated, and Vetëvendosje denied the claim.

Risk classification

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

The incident involved the coordinated use of AI-generated synthetic media to conduct a targeted political disinformation campaign aimed at manipulating public opinion and the electoral process.

Additional risk subdomains

• 3.1 False or misleading information: The AI system was used to generate a highly realistic but completely false image of a political meeting to mislead the public.

Causal factors

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

The incident was caused by human actors who intentionally generated and distributed a political deepfake post-deployment to influence an election.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The report describes the creation and dissemination of an AI-generated deepfake image, which falls under the category of AI-generated content requiring transparency obligations.

AI system and alleged parties

• AI system: Hive
• AI purpose: Deepfake Image Generation; Image Verification
• Behaviour type: Tool
• Alleged developer: Synthetic media generation technology developers, Image generation technology developers, Deepfake technology developers
• Alleged deployer: Information manipulation actors in Kosovo, Information manipulation actors, Gjithqka rreth Politikes (Facebook)
• Alleged harmed parties: Voters in Kosovo, Voters, Vetëvendosje, General public of Kosovo, General public, Epistemic integrity, Democratic integrity, Dejona Mihali

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 Minor
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Minor, indirect Minor
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Minor, indirect Minor
• 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.

Directly caused: An AI-generated deepfake image falsely showing Dejona Mihali meeting with Serbian President Vucic was created and shared on Facebook to discredit her.

Indirectly caused: The post spread false claims and accusations against Mihali, accusing her of working against Kosovo's interests in collusion with Serbia.

Inferred additional harm: The malicious content likely reached thousands of social media users, contributing to a toxic political discourse online during the election.

Democracy

Reported: The report explicitly describes the erosion of democratic norms, noting the image was shared during a snap-election campaign to influence the political process.

Directly caused: The deepfake was used to manipulate public opinion and discredit a political figure just a week before national elections.

Indirectly caused: The fabrication of political meetings undermines the integrity of the electoral process and democratic debate in Kosovo.

Inferred additional harm: Such incidents, if widespread, can erode public trust in democratic institutions, political parties, and the electoral system as a whole.

Epistemic

Reported: The report explicitly describes epistemic harm through the fabrication of a meeting and false quotes.

Directly caused: A fabricated image falsely depicting a meeting between Dejona Mihali and Aleksandar Vucic was published as real.

Indirectly caused: The post presented fake quotes from Mihali and Albin Kurti, distorting the truth of their political stances.

Inferred additional harm: The spread of highly realistic deepfakes pollutes the information ecosystem, making it harder for citizens to distinguish between real and fabricated political events.

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 1000

Potential causes

Technology

• Generative AI Accessibility: Easy access to AI tools enabled creation of realistic fake images.
• Absence of Platform Filters: Social media platforms failed to auto-detect and flag AI-generated deepfakes.

Data Inputs

• Target Image Availability: Publicly accessible photos of politicians were used to generate the deepfake.

Human Factors

• Intentional Deception: Malicious actors created the fake image to influence election outcomes.
• Political Polarization: High political tension made users receptive to spreading damaging disinformation.

Process and Methods

• Inadequate Page Moderation: The Facebook page published unverified fabricated content without checks.
• Reactive Fact-Checking: Fact-checkers analyzed the image only after it was widely distributed.

Regulatory Environment

• Weak Election AI Policies: Lack of strong regulatory frameworks to prevent deepfakes in elections.

Information quality

• Classification confidence: High
• Reason for confidence: The report provides clear, verified facts about the fabrication of the image, the specific political figures involved, the platform used for detection (Hive), and the electoral context. There is little ambiguity about what transpired.
• Ambiguities identified: The specific AI generator used to create the deepfake is not identified, only the detection tool (Hive) is named.
• Alternative interpretations: None. The evidence clearly points to a politically motivated disinformation attempt using synthetic media.

An AI-generated deepfake depicting a senior Kosovar official meeting with the Serbian president was disseminated on Facebook one week before national elections. The incident represents a targeted information operation aimed at manipulating democratic processes and electoral outcomes, highlighting the ongoing threat of synthetic media to national sovereignty and political stability.

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

Threat characteristics

• Imminence: Near-term. The campaign occurred one week prior to snap elections, presenting a rapidly developing threat to the integrity of the vote.
• Autonomy: Human-controlled. The AI system was used strictly as a tool to generate the fabricated image under direct human prompt and control.
• Novelty: Established threat. The use of synthetic media and deepfakes for political disinformation and character assassination is an established and widely observed threat.

Impact by dimension

• Physical security: Negligible. No physical systems, critical infrastructure, or human safety threats were involved in this digital disinformation incident.
• Information security: Substantial. The incident involved a targeted AI-generated deepfake campaign designed to mislead voters and manipulate public opinion during an active snap-election campaign.
• Sovereignty: Substantial. The deepfake targeted a senior political official with the intent to influence national election outcomes, representing a threat to democratic processes and electoral integrity.
• Economic security: Negligible. No strategic technology theft, financial system attacks, or economic manipulation occurred.
• Societal stability: Minor. The fabrication of political meetings contributes to the erosion of public trust in democratic institutions and pollutes the information ecosystem, though it did not cause civil unrest.