This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A deepfake scam impersonating Italian President Sergio Mattarella and Prime Minister Giorgia Meloni was used to promote a fraudulent cryptocurrency investment scheme. The videos, which were described as highly sophisticated, promised guaranteed income to lure victims into making initial deposits of €250. Italy's financial regulator, Consob, intervened by blocking the advertisements and websites, marking the first enforcement of new financial fraud regulations.
A deepfake scam allegedly impersonating Italian President Sergio Mattarella and Prime Minister Giorgia Meloni circulated on various platforms. It reportedly promoted a fraudulent cryptocurrency investment scheme. The deepfakes claimed the scheme was state-backed, promising €40,000 in guaranteed income. Victims were urged to invest €250, only to lose access to funds. Italy’s Consob blocked the deceptive ads and websites, marking the first enforcement of new financial fraud regulations.
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 deepfake videos generated false and misleading information regarding a state-backed guaranteed income program to deceive users.
The incident involves the use of AI-generated deepfakes of public figures to orchestrate a fraudulent cryptocurrency investment scam targeting consumers for financial 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 initiated by human scammers who intentionally deployed deepfake technology to defraud consumers.
Added categories peak at 2: Minor (Epistemic Harm, Direct)
Scammers deployed highly sophisticated deepfakes of Italy's President and Prime Minister to promote a fraudulent cryptocurrency scheme. While the immediate financial fraud was mitigated by regulatory intervention, the incident underscores the growing risk of high-level political impersonation for malicious purposes, posing minor threats to information security and public trust.
The AI system was used as a tool to generate synthetic video and audio (deepfakes) based on inputs provided by the human creators. The system itself did not exhibit autonomous decision-making or planning.
Limited Risk: The report describes the use of AI-generated content such as deepfakes, which are classified as limited risk under the EU AI Act and are subject to transparency obligations.
The AI system's primary function was to generate realistic synthetic videos of political figures, accompanied by synthetic voice generation to deliver the fraudulent message.
The report clearly outlines the nature of the scam, the specific public figures impersonated, the regulatory response by Consob, and the mechanics of the fraud. While the specific AI model is not named, this does not hinder the classification of the risk domain or causal factors.
None
Technology
Sophisticated Deepfake Generation
AI tools allowed creation of highly realistic fake videos of public figures.
Unregulated Hosting Platforms
Online platforms hosted and distributed fraudulent AI-generated videos.
Fake Trading Platform Software
Sites were designed to show simulated gains to deceive victims.
Data Inputs
Public Video and Audio Data
Abundant public footage of Giorgia Meloni was used to train the deepfake.
Human Factors
Trust in Public Figures
Victims trusted the message because it appeared to come from the Premier.
Financial Vulnerability
Promise of quick riches exploited victims' desire for financial improvement.
Lack of Deepfake Awareness
Users could not distinguish between real and AI-generated videos.
Process and Methods
Social Engineering Tactics
Criminals used phone follow-ups to guide victims through the deposit process.
Untraceable Financial Transfers
Wire transfers to non-existent foreign companies hid the money trail.
Delayed Ad Moderation
Ad networks failed to detect and block the fraudulent ads before publication.
Management
Ad Platform Profit Prioritization
Platforms accepted ad payments without verifying the legitimacy of the content.
Regulatory Environment
Lag in Regulatory Powers
Authorities previously lacked the power to quickly block unauthorized ads.
Cross-Border Jurisdictional Issues
Funds were sent to foreign entities, complicating recovery and prosecution.
