This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A disinformation campaign in Moldova utilized AI-generated synthetic media, including videos and photos, to depict Economic Development Minister Dumitru Alaiba in compromising situations. The campaign, attributed to pro-Kremlin actors, aimed to discredit the minister and influence the public ahead of presidential elections and a referendum on European Union integration. Government officials have denounced the content as fake, characterizing it as part of a broader psychological operation.
A purportedly AI-manipulated video and some photos circulated online depicting Moldova's Economic Development Minister, Dumitru Alaiba, in compromising situations as part of an alleged disinformation campaign by pro-Kremlin supporters, one that has been reported to rely on AI to generate fake content. Alaiba reportedly denounced the purported synthetic media as having been released to mislead the public and influence Moldova's elections by undermining his reputation and the government's pro-European Union stance.
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-generated video targeted a specific individual, Minister Alaiba, using humiliating synthetic imagery to damage his reputation.
The deepfake video was deployed as part of a coordinated, state-sponsored disinformation campaign aimed at manipulating public opinion and political processes in Moldova ahead of a crucial referendum and election.
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 who intentionally generated and distributed a deepfake video using post-deployment AI tools to conduct a political disinformation campaign.
Added categories peak at 2: Minor (Epistemic Harm, Direct)
Pro-Kremlin actors deployed an AI-generated deepfake video targeting Moldovan Economic Development Minister Dumitru Alaiba in a highly coordinated disinformation campaign. Occurring just before critical presidential elections and an EU integration referendum, the incident represents a substantial foreign interference attempt in domestic sovereignty and a notable escalation in information warfare capabilities.
The reports only mention that the video and images were artificially generated or fakes, without providing any details about the specific AI system used or its mode of operation.
Limited Risk: The incident involves AI-generated content such as deepfakes, which requires transparency obligations under the EU AI Act so that users are aware they are interacting with synthetic media.
The AI system was used to generate a synthetic video and accompanying photos depicting the minister in a fabricated scenario.
The reports provide consistent accounts from multiple sources confirming the existence of the fake video, the identity of the targeted minister, and the political context of the disinformation campaign. While the specific AI tool is not named, the classification of the incident as a deepfake-driven political disinformation campaign is highly certain.
None. All sources agree the video is a fabricated deepfake intended for political disruption.
Technology
Generative AI Deepfakes
AI used to create realistic fake video of the minister in a sauna.
Algorithmic Amplification
Social media algorithms prioritized and spread the sensational fake video.
Inauthentic Bot Networks
Automated accounts and bots amplified the reach of the deepfake.
Data Inputs
Targeted Facial Imagery
Minister's real facial features used to synthesize the deepfake video.
Fabricated Narrative Data
False textual claims of drug use paired with the generated video.
Human Factors
Cognitive Vulnerability
Public susceptibility to sensational and scandalous visual disinformation.
Malicious Actor Intent
Deliberate creation of deepfakes by adversaries to derail the referendum.
Confirmation Bias
Target audience's willingness to believe negative stories about politicians.
Process and Methods
Slow Content Moderation
Social platforms failed to rapidly detect and remove the deepfake video.
Inadequate Detection Tools
Lack of automated systems to instantly flag AI-generated disinformation.
Coordinated Inauthentic Behavior
Organized campaigns bypass standard platform reporting processes.
Management
Under-resourced Moderation
Social media companies allocated insufficient resources for Moldovan language.
Prioritizing Engagement
Platforms prioritizing traffic over proactive safety and verification.
Inadequate Risk Mitigation
Failure to plan for high-impact AI deepfakes during critical elections.
Regulatory Environment
Weak Platform Accountability
Lack of strict legal penalties for platforms hosting harmful deepfakes.
Geopolitical Exploitation
Lack of international legal frameworks to deter state-sponsored campaigns.
Election Integrity Gaps
Insufficient regulatory oversight on digital campaign materials.
AI system involved
An AI system was involved, but the reports do not identify it.
Identified by automated extraction; can differ from the alleged developers above.
