This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
U.S. Representative Derrick Van Orden used the Grok Imagine AI tool to generate and post deepfake videos of his congressional opponent, Rebecca Cooke, making false statements to damage her reputation for political gain.
Wisconsin U.S. Rep. Derrick Van Orden repeatedly posted Grok Imagine-generated videos depicting Democratic challenger Rebecca Cooke saying statements she did not make. Cooke called the clips deceptive and defamatory; Van Orden's campaign described them as obvious AI-generated satire. Cooke's attorneys later sent a cease-and-desist alleging the videos were intended to mislead voters and harm her reputation.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
Also classified as
The AI system was used to generate highly realistic but entirely fabricated statements and videos of a candidate, spreading false information to voters.
The incident involves a political candidate using AI-generated synthetic media to target, impersonate, and manipulate the public perception of a specific political opponent for personal and political gain.
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 incident was caused by a human political candidate intentionally using a deployed AI image and video generator to create and distribute deceptive deepfakes of his opponent.
Added categories peak at 2: Minor (Epistemic Harm, Direct)
A U.S. congressional candidate utilized a commercially available AI image generator to create and distribute defamatory deepfakes of his political opponent during an election campaign. This represents a minor domestic political incident using AI-enabled disinformation, posing negligible direct threats to national security, though reflecting broader societal challenges regarding synthetic media in democratic processes.
The AI system, Grok Imagine, was used as a tool by a human user to generate specific images and videos based on direct prompts, operating without independent planning or autonomous decision-making.
Limited Risk: The system involves AI-generated content (deepfakes) and chatbots, which are subject to specific transparency obligations under the Act, such as the automatic labels indicating they were made with 'Grok Imagine' mentioned in the report.
The primary function of the Grok Imagine tool in this incident was to generate synthetic, realistic video and image avatars of a real person speaking fabricated statements.
The report clearly identifies the political candidates involved, the specific AI platform used (Grok Imagine), the nature of the deepfake content, and the legal actions taken (cease-and-desist letter). There is no significant ambiguity regarding the facts of the deployment or the intent of the user.
None
Technology
Generative AI Video Tooling
Grok Imagine allowed the easy generation of realistic deepfake video content.
Lack of Deepfake Prevention
The AI platform lacked guards to block generating deepfakes of candidates.
Data Inputs
Target Reference Media
Publicly available video and audio of the candidate enabled realistic cloning.
Human Factors
Political Malice
A political opponent intentionally used deepfakes to damage a reputation.
Voter Susceptibility
Voters can be easily fooled or misled by realistic AI-generated videos.
Process and Methods
Automated Labeling Only
Platform relied on automatic labels, which failed to prevent dissemination.
Management
Permissive Platform Policies
Platform management allowed users to generate and share political deepfakes.
Regulatory Environment
Gaps in AI Statutory Laws
Wisconsin statutory protections lacked explicit, robust bans on political deepfakes.
Reliance on Common Law
Defamation common law had to fill gaps left by insufficient statutory rules.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
