The Arizona Agenda Produced a Deepfake of Kari Lake Advocating for the Publication Without Her Consent

The Arizona Agenda produced a deepfake video of Republican Senate candidate Kari Lake giving a testimonial about the publication with the seeming intention of educating the general public about the dangers of deepfakes in the coming election cycle. However, the Arizona Agenda appears not to have sought Lake's consent, prompting a cease-and-desist letter from her campaign.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.3 Fraud, scams, and targeted manipulation

The incident involved the unauthorized creation of a deepfake video of a political candidate to simulate a fake endorsement, which falls under impersonation and targeted manipulation of likeness.

Additional risk subdomains

• 3.1 False or misleading information: The deepfake video generated false information that misled some subscribers into believing the endorsement was real.

Causal factors

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

The incident was caused by humans (the journalists) intentionally using a post-deployment AI tool to generate a deepfake video of a political candidate for educational purposes.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Risk Level 3: Limited Risk. The incident involves AI-generated content (a deepfake), which falls under Limited Risk due to transparency obligations requiring users to be informed that they are interacting with or viewing AI-generated content.

AI system and alleged parties

• AI system: None named
• AI purpose: Deepfake Video Generation; Voice Generation
• Behaviour type: Tool
• Alleged developer: The Arizona Agenda, Hank Stephenson
• Alleged deployer: The Arizona Agenda, Hank Stephenson
• Alleged harmed parties: Kari Lake, Journalism, General public, Democracy

Harm severity

Highest direct severity in any category: Negligible. 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 Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 10000

Potential causes

Management

• Unintended Editorial Risk-Taking: Newsletter management deployed deepfakes despite risk of losing subscribers.
• Lack of Consent Protocols: Management failed to obtain permission before using candidate's likeness.

Technology

• Rapidly Improving AI Realism: AI-generated media is becoming nearly indistinguishable from real footage.
• High Accessibility of AI Tools: Free and easy-to-use tools allow anyone with a computer to generate fakes.
• Inherent Difficulty in Detection: Technical experts note that it is increasingly impossible to detect fakes.

Data Inputs

• Unauthorized Likeness Data: Publicly available video and voice data used to generate deepfakes.
• Lack of Authentication Standards: No cryptographic watermarks on official media to verify authenticity.

Human Factors

• Voter Unpreparedness: Electorate is unprepared for the influx of highly realistic fake media.
• Skepticism of Real Media: Public tends to write off objective, real videos as deepfakes.
• Susceptibility to Deception: Even sophisticated observers and paying readers were fooled by the fake.

Process and Methods

• Unconsented Demonstration Methods: Outlets create unauthorized deepfakes to demonstrate technology risks.
• Lack of Standardized Verification: Absence of third-party authentication methods for political communications.

Regulatory Environment

• Unregulated AI Technology: Lack of comprehensive laws governing deepfakes in political campaigns.
• Inadequate Legal Remedies: Campaigns must rely on cease-and-desist letters after the fake goes viral.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear, consistent, and detailed accounts of the creation, distribution, and reception of the deepfake video. The facts regarding the creators, the subject (Kari Lake), the platform (Arizona Agenda), and the resulting legal threat are well-documented across multiple sources.
• Ambiguities identified: The specific AI software or model used to generate the deepfake is not named in the reports.
• Alternative interpretations: None. The event is clearly an educational demonstration of deepfake technology that resulted in a cease-and-desist letter due to lack of consent.

A local political newsletter in Arizona created an unauthorized deepfake of Senate candidate Kari Lake as an educational demonstration. While intended to highlight AI risks, the incident caused minor voter confusion and a legal dispute, illustrating the growing challenge of synthetic media in democratic processes without posing a direct threat to national security.

• Overall national security impact: Minor
• Response level: Moderate
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Arizona Agenda

Threat characteristics

• Imminence: Long-term. The immediate incident is resolved, but it represents an ongoing strategic concern regarding the normalization of synthetic media in elections.
• Autonomy: Human-controlled. The AI was used strictly as a tool, with humans directing the creation, script, and deployment of the synthetic video.
• Novelty: Established threat. Deepfakes of political figures are an established threat, though this incident is notable for being conducted by a media outlet as an educational stunt.

Impact by dimension

• Physical security: Negligible. The incident involved a political deepfake video with no physical, kinetic, or critical infrastructure impact.
• Information security: Minor. While demonstrating deepfake capabilities, it was a localized educational stunt by a news outlet rather than a coordinated foreign information warfare operation.
• Sovereignty: Minor. The deepfake targeted a US Senate candidate, highlighting potential risks to electoral integrity, but had no systemic impact on government functions.
• Economic security: Negligible. No strategic technology theft or significant economic impact occurred, with financial consequences limited to a few cancelled newsletter subscriptions.
• Societal stability: Minor. The incident contributed to public confusion and erosion of trust in digital media, but did not result in large-scale civil unrest or systemic rights violations.