This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
AI-manipulated videos were used to target former Kosovo MP Paris Guri and LDK Youth Forum members ahead of the June 2026 snap elections. The videos, which were verified as AI-fabricated by the detection platform Hive, depicted youth members holding milk bottles and fabricated statements by Guri. These posts were identified as part of a disinformation effort during the election campaign period.
Purportedly AI-manipulated Facebook videos reportedly targeted former Kosovo MP Paris Guri and LDK Youth Forum members before Kosovo's June 7, 2026 snap elections. One video allegedly altered an original LDK image to show youth members holding milk bottles, while another reportedly fabricated Guri making a pro-Kurti statement.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to conduct large-scale disinformation campaigns, malicious surveillance, or targeted and sophisticated automated censorship and propaganda, with the aim of manipulating political processes, public opinion, and behavior.
Also classified as
The AI system was used to generate entirely fabricated statements and visual depictions that misled social media users.
The videos targeted specific individuals, including Paris Guri and LDK youth members, to humiliate them and manipulate public perception of their character.
The AI-manipulated videos were deployed as part of a coordinated political disinformation effort to influence public opinion and manipulate democratic processes ahead of snap elections.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by human actors intentionally using generative AI tools post-deployment to create and distribute deceptive political deepfakes.
Added categories peak at 2: Minor (Epistemic Harm, Direct)
AI-manipulated videos targeting political figures in Kosovo ahead of snap elections represent a localized political disinformation effort. While leveraging synthetic media to influence voter perception during sensitive democratic processes, the incident had minor national security impact and was quickly identified by fact-checking platforms.
The generative AI system was used as a tool by human actors to perform the specific function of generating and editing synthetic video and audio content based on direct inputs.
Limited Risk: The incident involves AI-generated content such as deepfakes, which are subject to specific transparency obligations under the EU AI Act.
The AI was used to generate synthetic video and voice content depicting real political figures saying and doing things they did not do.
The report provides clear, verified details about the targeted individuals, the platform used to detect the manipulation (Hive), and the political context. The facts are well-documented by the fact-checking organization Krypometer.
None
Technology
AI Video Manipulation Tools
Advanced AI tools allowed realistic editing of videos to depict false events.
High-Fidelity Voice Synthesis
AI synthesized the politician's voice to make false statements sound real.
Data Inputs
Misuse of Original Media
Authentic photos of LDK members were used as inputs for AI manipulation.
Human Factors
Malicious Intent of Publisher
An actor intentionally created deepfakes to manipulate public opinion.
Political Polarization
Electoral tension motivated creators to target opposition political figures.
Process and Methods
Lack of Platform Pre-Screening
Facebook allowed immediate publishing of AI-manipulated media without checks.
Delayed Detection and Verification
Fact-checking and HIVE analysis only occurred after the video went viral.
Regulatory Environment
Weak Social Media Regulation
Absence of strict legal penalties for distributing political deepfakes.
Lack of AI Election Safeguards
Insufficient regulatory frameworks to protect elections from AI manipulation.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
