Russian Disinformation Campaign Reportedly Used AI-Generated Posts and Videos to Target 2025 Moldovan Parliamentary Elections

Researchers and European officials reported that Russian operatives have been using purportedly AI-generated posts, videos, and websites to influence Moldova's September 2025 parliamentary elections. Networks of over 900 accounts across TikTok, Facebook, Instagram, Telegram, and YouTube have reportedly spread fabricated narratives, including reportedly misogynistic and false smears of President Maia Sandu.

Risk classification

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

The incident involves a coordinated, large-scale disinformation campaign using AI-generated content to influence the Moldovan parliamentary elections and public opinion.

Additional risk subdomains

• 1.2 Exposure to toxic content: The campaign featured misogynistic and homophobic smears targeting the female president.
• 3.2 Pollution of information ecosystem and loss of consensus reality: The massive influx of AI-generated fake posts, videos, and websites across multiple platforms polluted the local information ecosystem.

Causal factors

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

The risk was caused by the intentional actions of human Russian operatives deploying AI tools post-deployment to manipulate elections.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Risk Level 3: The report describes the use of AI-generated content, specifically fake videos and deepfakes, to spread disinformation, which falls under Limited Risk obligations.

AI system and alleged parties

• AI system: None named
• AI purpose: Social Media Content Generation; Deepfake Video Generation
• Behaviour type: Tool
• Alleged developer: Unknown generative AI developers
• Alleged deployer: Storm-1679, Storm-1516, Russian state-linked actors, Russian disinformation operators, Matryoshka, Influence operation groups
• Alleged harmed parties: National security and intelligence stakeholders, Moldovan voters, Moldovan democratic institutions, Maia Sandu, General public of Moldova, General public, European Union, Epistemic integrity, Electoral integrity

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Severe. 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 Severe, indirect Negligible
• Differential treatment: direct Minor, indirect Negligible
• Civil rights: direct Negligible, indirect Severe
• Democracy: direct Minor, indirect Substantial
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Minor
• Epistemic: direct Substantial, indirect Severe
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report explicitly describes toxic and malicious content created and spread directly by the incident.

Directly caused: Russian operatives directly created and spread AI-generated fake videos, posts, and websites containing misogynistic and homophobic smears targeting Moldovan leadership.

Inferred additional harm: Large numbers of social media users in Moldova were exposed to toxic and malicious content across TikTok, Telegram, Facebook, Instagram, and YouTube.

Differential treatment

Reported: The report explicitly describes differential treatment of female political leaders.

Directly caused: The disinformation campaign specifically targeted female leaders like President Maia Sandu with misogynistic narratives.

Inferred additional harm: This reinforces systemic gender-based discrimination and bias against women in political leadership positions.

Civil rights

Reported: The report explicitly describes efforts to undermine the democratic rights of the Moldovan electorate.

Indirectly caused: The campaign sought to undermine the Moldovan electorate's civil right to free and fair democratic elections through manipulation.

Democracy

Reported: The report explicitly describes erosion of democratic norms caused directly and indirectly by the incident.

Directly caused: The coordinated Russian campaign directly targeted the Moldovan parliamentary elections to manipulate public opinion and sway the vote.

Indirectly caused: The campaign stokes political and social divisions, threatening Moldova's national security, sovereignty, and democratic future.

Inferred additional harm: The integrity of the democratic process in Moldova is severely threatened by the scale of the AI-driven influence operations.

Psychological

Reported: The report explicitly describes targeted misogynistic and homophobic smear campaigns against President Maia Sandu.

Indirectly caused: The targeted misogynistic and homophobic smear campaigns likely caused distress and reputational harm to President Maia Sandu.

Inferred additional harm: Other female political figures and targeted individuals likely suffered psychological distress from these coordinated online attacks.

Epistemic

Reported: The report explicitly describes epistemic harm through the creation of fake posts, videos, and websites.

Directly caused: Operatives fabricated 39 distinct narratives and created fake videos and websites to spread false information.

Indirectly caused: The campaign allowed Russian actors to control the information environment, eroding shared truth and consensus reality.

Inferred additional harm: A significant portion of the Moldovan population of 2.4 million was likely misled or confused by the pervasive fake content.

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 100000

Potential causes

Management

• Slashed Democracy Support Funding: US cuts to foreign aid weakened Moldova's independent media and elections.
• Prioritizing Free Speech Over Safety: US administration prioritized free speech over regulating harmful AI content.

Technology

• Rapidly Evolving AI Tools: AI tools enabled easy creation of highly realistic fake videos and posts.
• Algorithmic Recommendations: Social media algorithms amplified fake AI content to increase engagement.

Data Inputs

• Fabricated Narratives as Input: Operatives fed targeted, fake narratives into AI to generate disinformation.
• Impersonation Data: Using real media branding like OK! magazine to train or prompt AI content.

Human Factors

• Exploitation of Social Biases: AI content targeted existing misogynistic and homophobic biases in society.
• Audience Vulnerability: Swaying a small percentage of divided voters can decide election outcomes.

Process and Methods

• Inadequate AI Content Detection: Platforms struggled to identify and remove coordinated AI disinformation.
• Evasion of Platform Moderation: Russian operatives used advanced tools to disguise account origins on apps.

Regulatory Environment

• Weakened Platform Regulations: US lobbying to weaken laws like the EU's Digital Services Act reduced oversight.
• Lack of Disinformation Policies: US government dismantled efforts and policies to combat foreign disinformation.

Information quality

• Classification confidence: High
• Reason for confidence: The report provides detailed accounts from multiple research organizations and government officials regarding the scale, platforms, and nature of the AI-driven disinformation campaign in Moldova. While the specific AI models are not named, the operational details of the campaign are well-documented.
• Ambiguities identified: The specific AI tools and models used to generate the fake videos and websites are not identified.

Russian state-sponsored operatives deployed an AI-driven disinformation campaign targeting Moldova's 2025 parliamentary elections and President Maia Sandu. Utilizing deepfakes and coordinated social media networks, the operation represents a severe foreign threat to Moldova's sovereignty, democratic processes, and information security.

• Overall national security impact: Severe
• Scope: Single nation
• Primary target: Moldova
• Alleged perpetrator: Russia

Threat characteristics

• Imminence: Near-term. Active and developing election interference campaign requiring ongoing monitoring and mitigation leading up to the vote.
• Autonomy: Human-controlled. AI systems were used as tools by human operatives to generate content, with humans controlling the dissemination and campaign strategy.
• Novelty: Evolved capability. Represents an evolved capability of existing foreign influence methods through the integration of highly realistic AI-generated content.

Impact by dimension

• Physical security: Negligible. No kinetic attacks, autonomous weapons deployment, or physical critical infrastructure manipulation reported.
• Information security: Severe. State-sponsored AI-driven disinformation campaign targeting national elections with coordinated networks and deepfakes to influence geopolitical alignment.
• Sovereignty: Severe. Foreign state-backed campaign directly targeting democratic parliamentary elections and the head of state to subvert national sovereignty and government alignment.
• Economic security: Negligible. No direct targeting of financial systems, strategic technology theft, or critical supply chains indicated.
• Societal stability: Substantial. Systematic use of toxic, misogynistic, and homophobic AI content to stoke social division and undermine democratic rights of the electorate.