This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Criminals in the Netherlands used AI-generated voice clips to impersonate bank officials in a sophisticated phone scam, defrauding approximately 170 victims of 380,000 euros.
Scammers allegedly used AI-generated bank-warning recordings to draw people in the Netherlands into calls with impersonators posing as bank staff. Victims were reportedly persuaded to install an app that relayed their bank cards' contactless data to accomplices at cash machines or payment terminals. Police reportedly documented about 170 victims and €380,000 in losses during three weeks of server monitoring.
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.
The perpetrators used AI-generated voice clips to impersonate bank officials and deceive specific victims for illegitimate financial benefit.
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 criminals who intentionally deployed AI voice generation technology to facilitate a financial scam.
Added categories peak at 2: Minor (Psychological Harm, Indirect)
Dutch criminals used AI-generated voice cloning to impersonate bank officials, defrauding 170 victims of 380,000 euros. While representing an evolution in cybercrime capabilities, the incident was localized, managed by standard law enforcement, and posed negligible threats to national security.
The AI system was used as a tool to generate static audio fragments ('tientallen verschillende geluidsfragmenten gemaakt met verschillende stemmen') which were played to victims during phone calls, performing a specific, well-defined function without autonomous decision-making or real-time interaction.
Limited Risk: The AI system was used to generate synthetic audio/deepfake voices to impersonate bank employees. Under the EU AI Act, AI systems that generate or manipulate image, audio or video content (deepfakes) are classified as limited risk and are subject to transparency obligations.
The AI was explicitly used to generate dozens of different voice clips mimicking bank and company representatives to deceive victims.
The reports from the Dutch police and AVROTROS are highly consistent, providing specific details on the number of victims (170), total financial damage (380,000 euros), the technical mechanism of the fraud (AI voice clips + NFC scanning app), and the arrest of the suspects. There are no major contradictions.
None. The reports clearly describe a coordinated cyber-enabled financial fraud campaign.
Technology
AI Voice Generation Tools
Allowed scammers to generate highly realistic and convincing voice recordings.
NFC Relay App
Allowed remote transmission of card data to accomplices at ATMs.
Automated IVR Spoofing
Used automated prompts ('press 1') to route victims to scammers.
Data Inputs
Stolen Customer Contact Info
Scammers used targeted phone numbers to contact potential victims.
NFC Chip Data Transmission
Unencrypted transmission of NFC card data over the rogue app.
Bank Branding Elements
Use of specific bank names and scenarios to train or script AI voices.
Human Factors
Susceptibility to Urgency
Victims acted under high stress and pressure to block their cards.
Overtrust in AI Voice Realism
Victims believed the highly realistic AI voices were actual bank employees.
Lack of Technical Literacy
Victims scanned physical cards against phones without knowing the risks.
Process and Methods
Insecure App Installation Process
Sideloading of unverified apps allowed malicious NFC relay tools.
Lack of Immediate Call Verification
No easy way for customers to instantly verify the validity of bank calls.
Scripted Social Engineering
Structured scripts maximized psychological pressure on victims.
Management
Inadequate Public Awareness
Public campaigns were insufficient to warn about AI voice scams.
Slow Initial Fraud Detection
Banks took weeks to identify and report the new NFC relay method.
Lack of Bank Coordination
Banks and police had delayed response until the ECTF meeting.
Regulatory Environment
Lack of AI Voice Regulations
No strict controls on the creation and deployment of voice cloning tools.
Permissive Sideloading Policies
Mobile operating systems allow installation of malicious non-store apps.
Weak App Store Controls
Lack of prevention of external malicious apps accessing NFC hardware.
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.
