This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
An advertising agency, DM9, used generative AI to manipulate footage of North Carolina State Senator DeAndrea Salvador from a 2018 TED Talk for a commercial campaign in Brazil. The manipulated video altered her voice, statements, and presentation slides to make it appear as though she were endorsing Whirlpool appliances. The senator filed a lawsuit alleging unauthorized use of her likeness, reputational harm, and emotional distress, leading to the withdrawal of the ad from the Cannes Lions International Festival of Creativity.
An advertising campaign created by DM9, a São Paulo–based subsidiary of Omnicom Group, for Whirlpool's Brazilian brand Consul reportedly used AI to alter footage from a 2018 TED Talk by North Carolina State Senator DeAndrea Salvador. The video reportedly modified her voice, statements, and slides to reference Brazilian energy data without her consent, appeared in an award-winning commercial, and later prompted a lawsuit alleging unauthorized use of her likeness and reputational harm.
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-manipulated ad generated and spread false statements about energy costs in São Paulo and fabricated a political figure's endorsement.
The ad agency used generative AI to manipulate the Senator's public likeness and voice to create a fake endorsement for commercial gain, which fits the boundary rule for targeted manipulation and fake endorsements.
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 the deliberate decision of humans at the advertising agency DM9 to use generative AI tools to manipulate the Senator's likeness for a commercial campaign.
Added categories peak at 2: Minor (Psychological Harm, Direct)
A Brazilian advertising agency used generative AI to deepfake a US State Senator's voice and likeness for a commercial appliance campaign. While representing a serious case of unauthorized synthetic media and likeness theft resulting in a civil lawsuit, the incident has negligible implications for national security.
The generative AI system was used as a tool by the ad agency's employees to perform specific, well-defined functions of video and voice manipulation based on direct human inputs.
Limited Risk: The incident involves the creation of AI-generated content (a deepfake video), which falls under the transparency and disclosure obligations of Risk Level 3.
The AI tools were primarily used to manipulate the Senator's video footage and alter her voice to make fabricated statements for a commercial advertisement.
The reports provide clear, consistent, and detailed accounts of the deepfake creation, the parties involved, the Senator's reaction, the legal actions, and the consequences at the Cannes Lions festival.
None
Technology
Generative AI Voice Cloning
AI was used to clone and change Salvador's voice to alter her statements.
Deepfake Video Generation
AI altered TED talk slides and video visuals without author's permission.
Data Inputs
Unauthorized Copyrighted Footage
DM9 used Salvador's 2018 TED Talk video without obtaining consent.
Manipulated Broadcast Data
Altered footage from a CNN Brazil broadcast was used in the commercial.
Human Factors
Unauthorised Employee AI Use
A small number of employees executed AI alterations without approval.
Lack of Creative Integrity
Creators fabricated statements to make it appear she endorsed Whirlpool.
Process and Methods
Inadequate Content Verification
Cannes Lions lacked verification to prevent material misrepresentation.
Lack of Agency Disclosure
DM9 submitted the ad to Cannes without disclosing the use of AI.
Management
Lack of AI Oversight
Whirlpool and Omnicom failed to verify the origin of agency-provided media.
Shadow AI Adoption
Employees used generative AI tools without management's knowledge.
Regulatory Environment
Absence of Deepfake Regulations
Lack of active laws regulating commercial deepfakes and AI synthetic media.
Failed Legislative Measures
NC bipartisan bill to regulate deepfakes failed to gain enough traction.
