Russian Center for Geopolitical Expertise Allegedly Used AI to Target U.S. Candidates with Disinformation in 2024 Election

The Moscow-based Center for Geopolitical Expertise (CGE) allegedly used AI to produce and spread deepfakes and disinformation targeting U.S. political candidates during the 2024 general election, aiming to sway public opinion and disrupt the electoral process. On December 31, 2024, the U.S. Treasury Department imposed sanctions against the CGE and its director, Valery Korovin, for their reported role in this interference.

Risk classification

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

The incident involves a coordinated, state-sponsored foreign influence campaign using generative AI to conduct large-scale disinformation operations aimed at manipulating the U.S. political process.

Additional risk subdomains

• 3.2 Pollution of information ecosystem and loss of consensus reality: The actors used AI to generate content across a network of over 100 fake news websites to create false corroboration and pollute the information ecosystem.

Causal factors

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

The disinformation campaign was intentionally orchestrated by human actors (CGE and GRU) using deployed generative AI tools to influence the 2024 U.S. election.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The reports describe the use of generative AI to create deepfakes and synthetic disinformation. According to the EU AI Act, AI systems that generate or manipulate image, audio or video content (deepfakes) pose a moderate risk and are subject to specific transparency obligations.

AI system and alleged parties

• AI system: generative AI tools
• AI purpose: Deepfake Video Generation; Social Media Content Generation
• Behaviour type: Tool
• Alleged developer: Valery Korovin, Main Intelligence Directorate (GRU), Center for Geopolitical Expertise (CGE), Aleksandr Dugin
• Alleged deployer: Valery Korovin, Main Intelligence Directorate (GRU), Center for Geopolitical Expertise (CGE), Aleksandr Dugin
• Alleged harmed parties: U.S. electorate, Electoral integrity, Democracy, 2024 U.S. elections candidates

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 Negligible
• 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 reports explicitly describe malicious content created and spread directly by the incident.

Directly caused: CGE used generative AI tools to quickly create disinformation and a manipulated deepfake video containing baseless accusations against a vice presidential candidate.

Inferred additional harm: It is highly likely that millions of social media users and website visitors were exposed to this malicious synthetic content, though the exact reach is not quantified.

Democracy

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

Directly caused: The campaign directly targeted the 2024 U.S. presidential election processes and institutions to stoke socio-political tensions and influence the electorate.

Indirectly caused: The operations sought to divide the American people and undermine public trust in democratic processes and institutions.

Inferred additional harm: The cumulative effect of such foreign influence operations is the long-term erosion of public trust in democratic elections and institutions.

Epistemic

Reported: The reports explicitly describe epistemic harm caused directly and indirectly by the incident.

Directly caused: CGE used generative AI to quickly create disinformation and a manipulated video producing baseless accusations against a vice presidential candidate.

Indirectly caused: The disinformation was distributed across a network of at least 100 websites designed to imitate legitimate news outlets to create false corroboration.

Inferred additional harm: The widespread dissemination of AI-generated fake news likely misled some voters and contributed to the pollution of the online information ecosystem.

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 150000000

Potential causes

Management

• GRU Financial Sponsorship: Funded server infrastructure and operations to enable AI disinformation.
• CGE Director Coordination: Oversaw the execution of AI-driven influence operations.
• State-Directed Mandate: Russian and Iranian leadership directed organizations to stoke discord.

Technology

• Generative AI Tools: Used to quickly create disinformation at scale for election interference.
• Deepfake Video Manipulation: Constructed baseless accusations targeting a vice presidential candidate.
• Self-Hosted AI Server: Hosted tools locally to bypass blocks by foreign web-hosting providers.

Data Inputs

• Manipulated Media Inputs: Used real videos as input to generate doctored deepfakes of candidates.
• Imitation News Content: AI was fed templates of legitimate news to mimic authentic outlets.

Human Factors

• U.S.-Based Facilitators: Helped build and maintain the AI-support server and pay rent.
• GRU Officer Direction: Directly coordinated and financed CGE personnel to run AI operations.
• Electorate Vulnerability: Susceptibility of U.S. voters to targeted digital disinformation.

Process and Methods

• Fake Website Networks: Used 100+ fake sites to distribute AI content and false corroboration.
• Obfuscation of Origin: Designed distribution methods to hide the Russian state origin of content.
• AI-Powered Botfarms: Used AI to automate nearly 1,000 fake social media personas.

Regulatory Environment

• Sovereign Enforcement Limits: Sanctions are reactive and struggle to stop active foreign AI operations.

Information quality

• Classification confidence: High
• Reason for confidence: The reports from multiple reputable sources (U.S. Treasury, major news outlets) consistently detail the sanctions, the entities involved (CGE, GRU), and the specific use of generative AI to create disinformation and deepfakes targeting the 2024 election. The role of the AI as a tool for content generation is clearly defined.
• Ambiguities identified: The specific generative AI models or platforms used by CGE are not named, and the exact reach or impact of the disinformation on the electorate is not quantified.
• Alternative interpretations: None. The incident is clearly a state-sponsored foreign influence campaign utilizing AI-generated content.

State-sponsored Russian and Iranian actors utilized generative AI to conduct a sophisticated disinformation campaign targeting the 2024 U.S. presidential election. By deploying deepfakes of candidates and leveraging self-hosted AI infrastructure to bypass controls, these adversaries sought to undermine democratic processes, posing a severe threat to national sovereignty and information security.

• Overall national security impact: Severe
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Russia and Iran

Threat characteristics

• Imminence: Long-term. Represents an ongoing strategic concern regarding foreign adversaries weaponizing AI for continuous and future election interference.
• Autonomy: Human-controlled. AI tools were used by human operators to generate content; the AI did not act autonomously or make independent decisions.
• Novelty: Evolved capability. While foreign election interference and deepfakes are known threats, the use of dedicated self-hosted AI servers to bypass hosting blocks represents an evolved capability.

Impact by dimension

• Physical security: Negligible. No physical infrastructure, kinetic attacks, or threats to physical safety were reported in connection with this incident.
• Information security: Severe. State-sponsored GRU operation utilized generative AI to create deepfakes of a VP candidate and run a network of 100 fake news sites to execute a sophisticated foreign influence campaign.
• Sovereignty: Severe. Direct attempt by Russian and Iranian state-backed entities to interfere in the 2024 U.S. presidential election and undermine public trust in democratic institutions.
• Economic security: Minor. While sanctions were applied, there was no direct threat to U.S. financial systems, strategic technology theft, or critical supply chain disruption reported.
• Societal stability: Substantial. Large-scale disinformation campaign designed to stoke socio-political tensions, polarize the electorate, and erode public trust in democratic processes.