AI-Generated Campaign Ad Reportedly Depicted Pennsylvania Governor Making Fabricated Statements About Sexual Harassment Scandals

Pennsylvania gubernatorial candidate Stacy Garrity's campaign reportedly released an AI-generated video impersonating Gov. Josh Shapiro and placing first-person statements about harassment scandals, intimidated accusers, and hush money in his mouth. The harassment allegations concerned a former aide. Shapiro's campaign denounced the ad as dishonest; a media scholar warned that its delayed AI disclosure could mislead viewers.

Risk classification

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

The AI system was used to generate a deepfake of a political candidate to influence public opinion and manipulate political processes during an election campaign.

Additional risk subdomains

• 3.1 False or misleading information: The ad generated discrete misleading content by using a synthetic likeness to falsely associate the governor with a sexual harassment complaint.

Causal factors

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

The risk was caused by the intentional decision of humans (the political campaign) to create and deploy a deepfake ad using a deployed AI tool.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The report describes an AI-generated campaign ad using a deepfake likeness, which falls under the category of AI-generated content such as deepfakes that require transparency disclosures.

AI system and alleged parties

• AI system: None named
• AI purpose: Deepfake Video Generation; Voice Generation
• Behaviour type: Tool
• Alleged developer: Synthetic video generation technology developers, Synthetic media generation technology developers, Synthetic audio generation technology developers, Deepfake technology developers
• Alleged deployer: Synthetic media creators, Stacy Garrity's gubernatorial campaign, Stacy Garrity, Public figures, Politicians in the United States, Politicians, political campaigns, Deepfake creators
• Alleged harmed parties: Public figures, Politicians in the United States, Politicians, Josh Shapiro, Epistemic integrity, Elected officials, 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 Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Minor, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report explicitly describes the creation of malicious synthetic content designed to attack a political opponent.

Directly caused: The campaign created and distributed an AI-generated deepfake ad targeting Governor Shapiro.

Democracy

Reported: The report explicitly describes concerns about the erosion of democratic norms through the manipulation of voter thoughts leading up to elections.

Directly caused: An AI-generated deepfake of a political candidate was used in an attack ad to mislead voters shortly before Election Day.

Inferred additional harm: The widespread, unregulated use of deepfakes in political campaigns could severely undermine public trust in democratic elections.

Epistemic

Reported: The report explicitly describes epistemic harm in the form of misleading political advertisements that manipulate voter beliefs.

Directly caused: The ad used a synthetic likeness of Governor Shapiro to deliver a misleading message, failing to disclose it was AI-generated for the first 45 seconds.

Inferred additional harm: Voters may have been misled into believing the governor himself was speaking or was directly involved in the sexual harassment complaint.

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 100000

Potential causes

Management

• Prioritizing Win Over Ethics: Campaign managers chose to use misleading AI to gain an advantage.

Technology

• Realistic Deepfake Generation: AI tools easily generate convincing likenesses of political figures.
• Delayed Disclaimer Timing: The ad hides the AI disclosure until 45 seconds into the video.

Data Inputs

• Unrestricted Training Data: Publicly available media of politicians is used to train AI models.

Human Factors

• Short Audience Attention Spans: Viewers absorb the misleading message before seeing the disclaimer.
• Campaign Intent to Mislead: Political actors willingly use deceptive AI to attack opponents.

Process and Methods

• No Mandatory Ad Labeling: No standardized process forces immediate labeling of AI content.

Regulatory Environment

• Absence of Federal Laws: No federal legislation restricts the use of deepfakes in elections.
• State Regulatory Gaps: Pennsylvania lacks active laws prohibiting deepfake political ads.

Information quality

• Classification confidence: High
• Reason for confidence: The report clearly details the creation and deployment of the AI-generated ad, the reactions of both campaigns, and the broader context of deepfakes in elections. There is little ambiguity about what occurred.
• Ambiguities identified: The specific AI tool or platform used to generate the deepfake is not mentioned.

An AI-generated political campaign ad used a deepfake likeness of Pennsylvania Governor Shapiro to deliver a misleading message shortly before Election Day. This incident represents a substantial threat to information security and democratic processes, reflecting an evolving trend of domestic political actors leveraging synthetic media to manipulate voters.

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

Threat characteristics

• Imminence: Near-term. The incident occurred less than two weeks before Election Day, representing a near-term threat to the integrity of the upcoming vote.
• Autonomy: Human-controlled. The AI system acted as a tool directly controlled and deployed by human campaign staff, who made all final decisions.
• Novelty: Evolved capability. Represents an evolved capability as deepfakes become increasingly accessible and integrated into mainstream political campaign strategies.

Impact by dimension

• Physical security: Negligible. No physical or kinetic threats, critical infrastructure manipulation, or safety risks were reported in connection with this incident.
• Information security: Substantial. The incident involves an AI-generated deepfake targeting a state governor to spread misleading information close to an election, showcasing AI-enhanced disinformation.
• Sovereignty: Substantial. The deployment of deepfakes in political campaigns directly threatens electoral integrity and democratic decision-making processes, though limited to a state-level race.
• Economic security: Negligible. No significant economic threats, strategic technology theft, or financial system manipulations were indicated.
• Societal stability: Minor. While eroding public trust in democratic processes, the incident did not cause mass civil unrest or systematic human rights violations.