Purportedly AI-Generated Videos Posted by Texas Lieutenant Governor's Campaign Allegedly Misrepresented Opponent's Tax Positions

Texas Lieutenant Governor Dan Patrick's campaign posted purportedly AI-generated videos depicting his opponent, state legislator Vikki Goodwin. She alleged the videos falsely portrayed her tax positions and were intended to deceive voters and harm her candidacy. Goodwin filed a police complaint, and the posts were later removed. Patrick's campaign called the videos a parody and denied violating election law.

Risk classification

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

The incident involves political campaigns using AI-generated deepfakes to conduct targeted disinformation and influence campaigns aimed at manipulating public opinion and election outcomes.

Additional risk subdomains

• 1.1 Unfair discrimination and misrepresentation: The AI ads misrepresented the political views of candidates, falsely depicting them saying things they did not say.
• 3.1 False or misleading information: The AI systems generated highly realistic but false videos of candidates to mislead voters.

Causal factors

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

The risk was caused by the intentional decisions of human political campaigns to create and distribute deceptive deepfake advertisements after the AI models were deployed.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The report describes the use of AI-generated content such as deepfakes, which under the EU AI Act are subject to transparency obligations to ensure users are informed they are interacting with or viewing AI-generated media.

AI system and alleged parties

• 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, Public figures, Politicians in the United States, Politicians, political campaigns, Deepfake creators, Dan Patrick's campaign, Dan Patrick
• Alleged harmed parties: Vikki Goodwin, Public figures, Politicians in the United States, Politicians, Political candidates targeted by deepfakes, 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 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 Substantial
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The reports describe the creation and distribution of deceptive AI-generated campaign ads (deepfakes) targeting political candidates.

Directly caused: Campaigns directly created and posted deepfake videos of candidates like Donna Goodwin and Sherrod Brown on social media platforms like X, reaching hundreds of thousands of viewers.

Inferred additional harm: It is likely that many other political campaigns utilized similar AI tools to generate deceptive content that went unreported.

Democracy

Reported: The reports explicitly describe the use of AI-generated deepfakes to deceive voters and influence election results, potentially violating state laws.

Directly caused: Political campaigns used AI to put false words into opponents' mouths, misrepresenting their policies to turn voters off and influence election outcomes.

Indirectly caused: The proliferation of 'slopaganda' and deceptive ads undermines the integrity of democratic elections and public trust in political processes.

Inferred additional harm: Continued unregulated use of AI deepfakes in elections is likely to further erode public trust in democratic institutions and the authenticity of political debate.

Epistemic

Reported: The reports describe the generation of false and misleading videos ('deepfakes') that fabricate candidates' statements and actions.

Directly caused: AI was used to generate realistic videos of candidates saying things they never said, such as Donna Goodwin advocating for tax increases.

Indirectly caused: The spread of these videos contributes to the pollution of the information ecosystem, making it harder for voters to distinguish real footage from synthetic parodies.

Inferred additional harm: The widespread adoption of AI-generated political ads will likely lead to a broader erosion of shared reality, where voters default to dismissing real evidence as deepfakes.

People affected

• Occurrences reported: 4
• People reportedly harmed: 3
• People reportedly exposed: 907000

Potential causes

Management

• Prioritizing Engagement Over Ethics: Campaign managers use deceptive AI to break through noise and win elections.
• Exploiting Legal Loopholes: Timing ad releases outside the 30-day window to evade legal penalties.

Technology

• Generative AI Deepfake Tools: Allows quick, inexpensive creation of highly lifelike, misleading ads.
• Lack of Built-in Watermarks: AI tools lack robust, mandatory indicators to easily identify synthetic media.

Data Inputs

• Unauthorized Candidate Likeness: Using real candidates' voices and faces to train or generate fake footage.
• Unauthorized Trademark Use: Using university logos and symbols without authorization in AI generation.

Human Factors

• Intent to Mislead Voters: Campaigns deliberately using AI to put false words in opponents' mouths.
• Voter Susceptibility to Deepfakes: Voters struggle to distinguish lifelike AI-generated ads from real footage.

Process and Methods

• Inadequate Platform Moderation: Social media platforms lack proactive filtering to detect and block deepfakes.
• Absence of AI Disclosures: Campaigns publishing synthetic media without clear, prominent labeling.

Regulatory Environment

• Lack of Federal AI Regulation: No federal laws govern the use of deceptive AI in political campaigns.
• Flawed State Deepfake Laws: Texas law has a narrow 30-day window and hard-to-prove intent requirements.
• Failed Legislative Updates: Bills to close regulatory loopholes failed due to free speech objections.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear, consistent, and detailed accounts of multiple political campaigns using AI-generated deepfakes to target opponents. The facts regarding the creation of the ads, the candidates targeted, the public reactions, and the legal complaints are well-documented across multiple reputable news sources.
• Ambiguities identified: There is some ambiguity regarding whether the creators of the ads intended them as obvious parodies (as claimed by Patrick's campaign) or as deliberate deceptions (as claimed by Goodwin).
• Alternative interpretations: The ads could be interpreted as protected political parody/satire rather than malicious disinformation campaigns.

Political campaigns in multiple US states utilized generative AI to create highly realistic deepfakes of opponents, aiming to mislead voters. This represents a substantial threat to democratic integrity and information security, signaling an evolved capability in domestic election interference.

• Overall national security impact: Substantial
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Domestic political campaigns and super PACs

Threat characteristics

• Imminence: Near-term. Deepfakes deployed within active election windows present a near-term threat to voting integrity requiring rapid response before election days.
• Autonomy: Human-controlled. AI tools were strictly used by human operators to generate specific content; the systems did not act or deploy ads autonomously.
• Novelty: Evolved capability. While political dirty tricks and basic deepfakes have existed, the scale, low cost, and realism represent a significant evolution in capability.

Impact by dimension

• Physical security: Negligible. No physical security threats, kinetic attacks, or critical infrastructure disruptions were reported in connection with these political deepfakes.
• Information security: Substantial. Represents a notable escalation in AI-enhanced domestic disinformation, using realistic deepfakes to target political candidates and mislead voters at scale.
• Sovereignty: Substantial. Directly targets the integrity of democratic elections and core government processes by manipulating voter perception through deceptive synthetic media.
• Economic security: Negligible. No direct threats to financial systems, strategic technology theft, or supply chain disruptions were identified in this incident.
• Societal stability: Minor. Contributes to the erosion of public trust and epistemic security, though currently managed through state laws and platform-level content removals.