This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Russian operatives, identified by Microsoft as the group Storm-1516, produced a fake video featuring a synthetically generated voice named 'Olesya' to spread disinformation regarding the U.S. election. The video falsely claimed that a Kyiv-based troll farm was working to influence American voters. U.S. intelligence confirmed the voice was AI-generated, highlighting the use of synthetic media as a tool for foreign influence operations intended to erode trust in democratic institutions.
Russian operatives used AI to create a fake video and voice of "Olesya," a supposed troll in Kyiv, falsely claiming involvement in U.S. elections to support President Biden. U.S. intelligence confirmed the voice was AI-generated. This disinformation campaign aimed to mislead voters, erode trust in democratic institutions, and influence the 2024 election. The incident involved the group Storm-1516, individuals linked to Valery Korovin, and potential veterans of the Internet Research Agency.
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 false and misleading narrative content to deceive voters.
The incident involves Russian state-aligned actors using synthetic voice generation to conduct a coordinated disinformation campaign aimed at manipulating the U.S. political process and public opinion.
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 disinformation campaign was intentionally designed and executed by human actors using a deployed AI voice generation tool to influence the U.S. election.
Added categories peak at 4: Severe (Epistemic Harm, Indirect)
Russian-aligned group Storm-1516 utilized AI-generated synthetic voice technology to fabricate a disinformation video targeting the U.S. presidential election and relations with Ukraine. While the video had limited public traction, the incident underscores the evolving threat of AI-enabled foreign influence operations aimed at undermining democratic institutions and sovereign government credibility.
The AI system functioned as a Tool, performing the specific, well-defined task of generating a synthetic voice based on direct inputs to produce a fake video narrative.
Risk Level 3. Limited Risk: The report describes the use of AI-generated content, specifically a synthetic voice in a deepfake video, which falls under the transparency obligations of Limited Risk systems.
The AI system was specifically used to generate a synthetic voice ('Olesya') to narrate a fake video designed for social media circulation.
The report clearly details the nature of the incident, the actors involved (Storm-1516), the specific AI technology used (synthetic voice generation), and the intended political impact. There is high consensus among U.S. intelligence and Microsoft researchers regarding the facts.
None. The evidence pointing to a coordinated Russian disinformation campaign using synthetic media is robust and uncontested in the text.
Technology
Synthetic Voice Generation
AI-generated voice was used to create realistic narration for the fake video.
AI Accent Obscuration
AI tools were used in propaganda videos to obscure foreign accents.
Human Factors
Susceptibility to Realistic Audio
Audiences are highly susceptible to realistic, AI-altered audio messages.
Process and Methods
Slow Dissemination of Verification
Agencies struggle to quickly expose and publicize AI-manipulated media.
Management
Targeted Disinformation Campaigns
Foreign threat actors systematically manage AI tools to target US voters.
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.
