Purported Deepfake Videos Allegedly Impersonated a Former Student to Falsely Accuse Texas U.S. Senate Candidate James Talarico of Sexual Assault

Two purported deepfake videos circulated on X using former student Zyon Lewis's identity to allegedly falsely accuse Texas U.S. Senate candidate James Talarico of sexual assault. Lewis said the person depicted was not him and denied the allegation. Researchers linked the videos to the Russia-aligned Storm-1516 influence operation. The original account was suspended, but copies continued to circulate.

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-aligned foreign influence operation using AI-generated deepfakes to target a political candidate and manipulate public opinion during an election.

Additional risk subdomains

• 3.1 False or misleading information: The AI system was used to generate highly damaging, completely fabricated claims of sexual assault to mislead voters.
• 4.3 Fraud, scams, and targeted manipulation: The deepfake videos impersonated a specific, real individual (a former student) without his consent to damage the candidate's reputation.

Causal factors

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

The incident was driven by human actors (the Storm-1516 group) who intentionally used deployed AI deepfake tools to create and spread political disinformation.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The AI system was used to generate deepfakes (AI-generated content). According to the EU AI Act, providers of AI systems that generate or manipulate image, audio or video content (deepfakes) must disclose that the content has been artificially generated or manipulated.

AI system and alleged parties

• AI system: unspecified
• AI purpose: Deepfake Video Generation; Voice Generation
• Behaviour type: Tool
• Alleged developer: Synthetic video generation technology developers, Synthetic media generation technology developers, Deepfake technology developers
• Alleged deployer: Synthetic media creators, Storm-1516, Information manipulation actors in Russia, Information manipulation actors, Deepfake creators
• Alleged harmed parties: Zyon Lewis, Voters in the United States, Voters, Targets of online impersonation, Public figures, Privacy, Politicians in the United States, Politicians, political candidates, James Talarico, Epistemic integrity, Democratic integrity

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 Minor
• 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: Yes, the report describes the creation and spread of deepfake videos containing false accusations of sexual assault.

Directly caused: The Russian disinformation campaign Storm-1516 generated two deepfake videos accusing a political candidate of sexual assault, which were posted on X.

Indirectly caused: The videos were amplified by known Storm-1516 amplifiers, conservative X networks, and a QAnon Facebook account, reaching an unspecified but likely large online audience.

Democracy

Reported: Yes, the reports describe foreign interference in a U.S. Senate election using AI-generated disinformation.

Directly caused: The Russian propaganda campaign Storm-1516 used AI-generated deepfakes to target Democratic U.S. Senate candidate James Talarico in an attempt to influence the election.

Indirectly caused: The campaign sought to shape public discourse, exacerbate extreme partisanship, and influence voter behavior in the Texas Senate race.

Epistemic

Reported: Yes, the report describes the creation and spread of fabricated video evidence to mislead the public.

Directly caused: Two deepfake videos were created and spread on social media, falsely depicting a former student accusing a candidate of sexual assault.

Indirectly caused: The false narrative was amplified across multiple social media platforms, including X, Facebook, and TikTok, polluting the information ecosystem during an election campaign.

People affected

• Occurrences reported: 1
• People reportedly harmed: 2
• People reportedly exposed: 125400

Potential causes

Management

• Engagement-Driven Algorithms: Platform management prioritized viral reach over information integrity.
• Poor Election Risk Planning: Failure to allocate sufficient resources to counter foreign AI campaigns.

Technology

• Accessible Deepfake Tools: AI tools enabled synthesis of realistic video and voice to mimic a real person.
• Flawed AI Generation Artifacts: Unnatural eye movements and stagnant expressions revealed the video was fake.
• Weak Automated Detection: Platforms lacked automated tools to block deepfakes during upload.

Data Inputs

• Harvested Personal Data: Public Facebook data of a real student was used to craft the fake persona.
• Missing Provenance Metadata: Lack of cryptographic watermarks allowed anonymous AI video distribution.

Human Factors

• Partisan Confirmation Bias: Conservative accounts reshared the video without verifying its authenticity.
• Malicious Foreign Actors: Russian group Storm-1516 intentionally fabricated the sexual assault claim.
• Impersonation Vulnerability: A real individual's identity was stolen to make the deepfake credible.

Process and Methods

• Coordinated Amplification: Networks of bots and paid influencers quickly spread the deepfake video.
• Reactive Platform Moderation: X suspended the account only after the video had already gone viral.
• Weak Identity Verification: An account registered in Romania easily impersonated a Texas resident.

Regulatory Environment

• Lack of AI Election Laws: No federal regulations prohibiting AI-generated political disinformation.
• Weak Platform Liability: Social media platforms face few legal penalties for hosting deepfakes.

Information quality

• Classification confidence: High
• Reason for confidence: The reports from Graphika and PolitiFact provide clear, consistent, and detailed evidence of a coordinated foreign influence operation using AI-generated deepfakes to target a political candidate. The roles of the actors, the platforms used, and the fact-checking process are well-documented, leaving little ambiguity about the nature of the incident.
• Ambiguities identified: The specific AI tools or models used to generate the deepfake videos are not identified.

The incident involves a coordinated Russian state-aligned information operation targeting a U.S. Senate candidate with AI-generated deepfakes to influence an election. While the AI generation was low quality and quickly debunked, it represents a substantial threat to information security and national sovereignty by attempting to manipulate democratic electoral processes.

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

Threat characteristics

• Imminence: Near-term. The campaign occurred during an active election cycle, presenting a near-term threat to the integrity of the voting process.
• Autonomy: Human-controlled. AI tools were used strictly as a tool by human operators to generate the video and audio content, with no autonomous decision-making.
• Novelty: Evolved capability. Represents an evolved capability of foreign actors utilizing synthetic media to target down-ballot candidates with highly damaging personal accusations.

Impact by dimension

• Physical security: Negligible. No physical threat, kinetic targeting, or critical infrastructure compromise was reported in connection with this digital disinformation campaign.
• Information security: Substantial. Coordinated Russian state-aligned influence operation (Storm-1516) used AI-generated deepfakes to conduct information warfare targeting a U.S. political candidate.
• Sovereignty: Substantial. Active foreign interference in a U.S. Senate election represents an attempt by a foreign adversary to manipulate sovereign democratic processes and government elections.
• Economic security: Negligible. The incident did not target financial systems, strategic industries, or critical supply chains, resulting in negligible economic impact.
• Societal stability: Minor. The campaign aimed to erode trust and manipulate public opinion, but impact was limited to a specific candidate and did not cause mass civil unrest.