This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Fraudsters utilized generative AI tools to create a sophisticated clone of the brokerage firm Exante, including a replica trading interface and fake identification documents. This clone was used to deceive a U.S. investor into transferring funds to a fraudulent JPMorgan Chase account. The incident highlights the growing use of AI in large-scale financial fraud and identity theft.
Scammers used AI tools to clone the broker Exante and defraud at least one U.S. victim by registering a JPMorgan Chase account and replicating Exante’s trading interface. AI-generated fake documents, deepfakes, and cloned websites enabled the scheme. Exante, which does not operate in the U.S., confirmed the fraud and filed reports with multiple U.S. agencies.
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
Fraudsters used AI-powered deepfakes to bypass biometric liveness detection tools used by fintech companies.
Fraudsters used generative AI to clone a financial platform and deceive a U.S. investor into transferring funds, which is a clear case of AI-facilitated fraud and targeted manipulation 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 caused by human fraudsters intentionally deploying and using generative AI tools to commit financial fraud after the models were trained and released.
Added categories peak at 3: Substantial (Epistemic Harm, Inferred)
Criminal actors utilized generative AI to execute highly sophisticated financial fraud, including cloning a brokerage platform and bypassing biometric identity verification. While representing a significant evolution in cybercrime and synthetic media capabilities, the incident constitutes a localized financial and criminal threat rather than a direct challenge to national security.
The AI systems, such as ChatGPT and deepfake generators, were used as tools by fraudsters to perform specific functions like generating fake passport images, cloning voices, and writing website code based on direct inputs.
Risk Level 3: Limited Risk. The report describes the use of AI-generated content such as deepfakes and cloned websites to deceive individuals, which falls under the category of deepfakes and interactive AI systems requiring transparency.
The fraudsters primarily used generative AI to create deepfake images (such as fake passports) and cloned voices to bypass security and deceive victims.
The reports provide detailed, consistent accounts of how generative AI tools are being used to facilitate financial fraud, specifically detailing the Exante clone case with statements from Exante's Head of Compliance and regulatory warnings.
None. The incident is clearly a case of AI-facilitated financial fraud.
Technology
AI Deepfake Generation
Generative algorithms clone voices and faces to bypass verification.
Liveness Detection Bypass
Pretrained AI tools allow scammers to bypass biometric authentication.
Rapid Website Cloning
Generative AI creates highly realistic clone websites in seconds.
Data Inputs
Minimal Data Requirements
Scammers need only one minute of audio and one photo to clone targets.
Algorithmic Profile Matching
Algorithms analyze data to target specific vulnerable user profiles.
Fake Document Generation
AI tools generate highly realistic fake passport and ID images.
Human Factors
Trust in Video Conferencing
Staff members trust realistic live-video clones of senior management.
Susceptibility to Urgency
Urgent fake fraud warnings manipulate customers into sharing data.
Cognitive Overload Exploitation
Scammers target victims during peak periods when monitoring is harder.
Process and Methods
Flawed Liveness Verification
Traditional liveness checks fail against sophisticated AI deepfakes.
Inadequate KYC Standards
Know-Your-Customer processes fail to detect synthetic identities.
Lack of Behavioral Analytics
Firms fail to track behavioral anomalies like mouse movements.
Management
Inadequate Risk Assessments
Financial firms fail to conduct thorough checks during account creation.
Slow Threat Response
Management fails to quickly report and block cloned platforms.
Regulatory Environment
Lagging Regulatory Action
Regulators have not fully addressed AI-driven cyber threats.
Delayed DeFi Regulations
New rules impacting DeFi will not come into effect until 2027.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
