Kremlin-Linked Entities Allegedly Using Generative AI to Spread Russian Disinformation in Latin America

Moscow-based tech firms and an industry association with links to the Kremlin are allegedly using generative AI to spread Russian disinformation in countries throughout Central America and South America. According to the U.S. Department of State, the Russian companies rely on local writers to compose stories which are then amplified across social media using artificial intelligence chatbots.

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 disinformation campaign targeting Latin American public opinion and political alignment using AI chatbots for amplification.

Additional risk subdomains

• 3.1 False or misleading information: The campaign actively generated and spread false claims, such as the fabricated story about the West looting religious relics from a Kyiv monastery.

Causal factors

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

The disinformation campaign is intentionally designed and coordinated by human actors associated with the Russian government, who plan to leverage AI chatbots to amplify their propaganda.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The report describes the planned use of 'artificial intelligence chatbots to amplify the articles on social media', which falls under the category of chatbots and AI-generated content requiring transparency.

AI system and alleged parties

• AI system: artificial intelligence chatbots
• AI purpose: Social Media Content Generation; Automated Content Curation
• Behaviour type: unknown
• Alleged developer: unknown
• Alleged deployer: Structura National Technologies, Social Design Agency, Oleg Yasinsky, Oleg Yasinskiy, Nikolay Tupikin, Institute for Internet Development, Ilya Gambashidze, Andrey Perla
• Alleged harmed parties: Ukraine, News media in Latin America, Journalistic integrity, General public

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 report explicitly describes the creation and spread of malicious propaganda and disinformation designed to undermine support for Ukraine.

Directly caused: The campaign directly spread false claims, such as the allegation that the West was looting religious relics and art from a monastery in Kyiv.

Inferred additional harm: It is likely that the campaign created and spread numerous other pro-Kremlin and anti-U.S./anti-NATO articles and social media posts across Latin America.

Democracy

Reported: The report explicitly describes attempts to erode democratic norms by interfering in political processes and eroding public faith in elections.

Directly caused: The campaign surreptitiously exploited the openness of Latin America's media environment to manipulate public opinion and weaken international support for Ukraine.

Indirectly caused: The U.S. intelligence assessment noted that Moscow uses such campaigns to erode public faith in the integrity of democratic elections.

Inferred additional harm: The campaign likely weakened democratic discourse and social cohesion in the targeted Latin American nations by injecting polarizing state-sponsored propaganda.

Epistemic

Reported: The report explicitly describes epistemic harm through the laundering of propaganda and fake news to make it appear organic to local audiences.

Directly caused: The campaign published false claims, such as a fabricated story about the West looting religious relics from a Kyiv monastery, to mislead audiences.

Indirectly caused: The use of AI chatbots to amplify these articles on social media further pollutes the information ecosystem with deceptive content.

Inferred additional harm: The widespread dissemination of localized Russian propaganda likely led to increased belief in false narratives and a degraded shared understanding of the conflict in Ukraine among Latin American audiences.

People affected

• Occurrences reported: 1

Potential causes

Management

• State-Funded Campaigning: Russian government financing and directing the influence operations.
• Coordinated Influence Firms: SDA and Structura managing the campaign execution under state direction.
• Strategic Narrative Control: Ensuring campaign themes stay strictly on topics of priority to the Kremlin.

Technology

• AI Chatbots for Amplification: AI chatbots are planned to amplify disinformation articles on social media.
• Deep Technical Capabilities: Influence firms exploit open information environments with technical tools.
• Fake News Site Spoofing: Creating fake articles purporting to be from actual trusted news organizations.

Data Inputs

• Laundered Russian Claims: False claims and propaganda are used as the primary content inputs.
• Linguistic Customization: Spanish and Portuguese translations make foreign propaganda feel organic.
• Targeting Divisive Discourse: Amplifying preexisting popular or divisive discourse to penetrate environments.

Human Factors

• Co-opted Local Influencers: Local writers and public figures are leveraged to publish pro-Kremlin views.
• Cultivated Journalists: Journalists in target countries review and edit content to localized standards.

Process and Methods

• Information Laundering: Sourcing articles locally to hide Russian government involvement.
• Coordinated Social Amplification: Using networks of proxy websites and social media to spread stories.
• Prebunking Limitations: Debunking narratives in advance is limited and hard to do in real time.

Regulatory Environment

• Open Information Ecosystem: Exploiting the openness of Latin American media and information systems.
• Lack of Foreign Influence Regulation: Inability to prevent covert foreign state-sponsored information campaigns.

Information quality

• Classification confidence: Medium
• Reason for confidence: The reports provide extensive details on the actors, mechanics, and goals of the Russian disinformation campaign. However, the AI component (the use of chatbots for amplification) is only mentioned as an intended tactic with no technical specifications, deployment metrics, or direct evidence of active execution, which limits our certainty regarding the AI-specific aspects of the incident.
• Ambiguities identified: The exact role, scale, and deployment status of the AI chatbots in the amplification process remain unclear and are presented as an intended plan rather than an active, documented deployment.
• Alternative interpretations: The incident could be interpreted primarily as a traditional human-driven influence operation with minor or speculative AI involvement, rather than an AI-driven safety incident.
• Missing information: Technical details of the AI chatbots, evidence of their active deployment, the specific social media platforms targeted by the AI, and metrics on the reach or impact of the AI-amplified content.

A state-sponsored Russian disinformation campaign targeting Latin America planned to use AI chatbots to amplify anti-US, anti-NATO, and anti-Ukraine narratives. While the campaign represents a notable information warfare operation (substantial impact), the AI component remains a tool for human-controlled amplification rather than an autonomous threat.

• Overall national security impact: Substantial
• Scope: Multiple nations
• Primary target: Latin America
• Other affected: United States, Ukraine, NATO countries
• Alleged perpetrator: Russia

Threat characteristics

• Imminence: Long-term. Represents an ongoing strategic information warfare concern rather than an active, immediate crisis requiring a 72-hour response.
• Autonomy: Human-controlled. AI chatbots are used as tools to assist and amplify content that is created, reviewed, and curated by human actors.
• Novelty: Evolved capability. State-sponsored disinformation is an established threat, but integrating AI chatbots to scale and automate the amplification of localized content represents an evolved capability.

Impact by dimension

• Physical security: Negligible. No physical threats, kinetic attacks, or critical infrastructure manipulation were reported.
• Information security: Substantial. Russian state-sponsored campaign intending to use AI chatbots to amplify anti-US and anti-NATO disinformation across Latin America, representing a notable information warfare threat.
• Sovereignty: Minor. Aims to manipulate public opinion and erode international support for Ukraine, but does not directly compromise core sovereign government decision-making.
• Economic security: Negligible. No significant economic disruption, financial system attacks, or strategic technology theft were reported.
• Societal stability: Minor. Attempts to manipulate public opinion at scale, which could degrade social cohesion, but there is no evidence of direct civil unrest or systemic human rights violations.