This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A fraudulent campaign used AI-generated deepfake videos and voice clones of celebrities to market honey-based supplements as cures for Alzheimer's disease. Victims were lured into making purchases that often led to unauthorized recurring charges and financial loss. The Better Business Bureau issued warnings about the scam, noting that the products have no proven medical benefit.
Beginning around October 2025, a campaign reportedly used AI-generated videos and cloned voices of well-known public figures and celebrities such as Bill Gates, Sanjay Gupta, Anderson Cooper, Bruce Willis, and others to market honey-based supplements as treatments for Alzheimer's disease and dementia. Consumers reportedly said modest purchases sometimes led to continuing charges or unwanted shipments, while cancellation and refund requests were difficult to resolve.
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 deepfakes spread false medical claims that honey can cure Alzheimer's, misleading consumers about health treatments.
The incident involves scammers using AI-generated deepfakes of celebrities to deceive individuals into purchasing fake medical cures, resulting in financial fraud.
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 driven by human scammers who intentionally deployed AI-generated deepfakes post-deployment to defraud consumers.
Added categories peak at 3: Substantial (Epistemic Harm, Inferred)
This incident involves a sophisticated consumer fraud campaign utilizing AI-generated deepfakes and voice clones of prominent figures to sell fraudulent medical treatments. While it highlights the growing ease of generating highly convincing synthetic media for malicious purposes, its national security impact is minor, representing localized criminal activity rather than a threat to state security.
The AI system was used to generate deepfake videos and voice clones of celebrities based on direct inputs. The reports describe the technology as 'powerful AI tools to create realistic videos' that perform specific generation functions without autonomous planning or agency.
Risk Level 3. Limited Risk: The report describes 'AI-generated content such as deepfakes' used to promote products. Under the EU AI Act, deepfakes are classified as limited risk, requiring transparency obligations to inform users they are interacting with AI-generated media.
The AI system was primarily used to create realistic deepfake videos and voice clones of celebrities to endorse a fake cure.
The reports provide clear, consistent accounts of the scam, including a detailed first-hand victim report with specific financial losses ($2,003) and a warning from the Better Business Bureau. The role of AI (deepfakes and voice clones) is explicitly stated, and the fraudulent nature of the campaign is verified by the BBB, the Alzheimer's Association, and Bill Gates's office.
None. The evidence clearly points to a coordinated financial scam using synthetic media.
Technology
AI Deepfake Video Generation
Realistic AI-generated videos of celebrities were used to deceive victims.
Unregulated AI Video Tools
Powerful AI tools allowed scammers to easily synthesize celebrity voices.
Data Inputs
Celebrity Likeness Training Data
AI models trained on public videos of celebrities to create deepfakes.
Falsified Health Questionnaires
Scammers collected personal health data to customize high-pressure sales.
Human Factors
Trust in Public Figures
Victims trusted the deepfaked endorsements of familiar celebrities.
Fear of Cognitive Decline
Exploitation of fears related to Alzheimer's and dementia.
Low Digital Media Literacy
Difficulty distinguishing between genuine videos and AI deepfakes.
Process and Methods
Weak Ad Moderation Processes
Social media platforms failed to detect and block AI-generated scam ads.
Deceptive Subscription Billing
Scammers used hidden recurring charges and complex cancellation flows.
High-Pressure Sales Funnel
Urgent warnings and follow-up calls forced quick purchases.
Management
Platform Profit Over Safety
Social media platforms prioritized ad revenue over strict safety vetting.
Inadequate Risk Assessment
Failure of platforms to anticipate AI-enabled fraud vectors.
Regulatory Environment
Lack of Deepfake Regulations
Inadequate legal frameworks to prevent the malicious use of AI deepfakes.
Weak Online Advertising Oversight
Regulatory gaps in monitoring truth-in-advertising on social platforms.
AI system involved
An AI system was involved, but the reports do not identify it.
Identified by automated extraction; can differ from the alleged developers above.
