AI-Driven Mule Account Identification & Alert System-Kaohsiung City Police Department
★2025 Interior Hackathon AI Application Competition | Bronze Award★
Intercepting Fraud at the Source: New Bank Account Creation
As fraud attempts proliferate, Kaohsiung is leveraging AI to detect "mule accounts" (accounts used for money
laundering). By building a big data model trained on known fraud characteristics, the system identifies suspicious
patterns in real-time. This addresses the time lag inherent in manual reviews, allowing financial institutions to blockfraudulent cash flows and money laundering instantly, effectively safeguarding citizens' assets.
Identifying Fraudulent Patterns
The system integrates diverse data points, including login IPs during account creation, applicant residency,
and historical profiles of known mule accounts. Using Logistic Regression AI, the system masters the behavioral
indicators of potential fraud. This eliminates risk management blind spots, creating multiple benefits: early
warnings, blocked transactions, and reinforced anti-money laundering (AML) protocols.
Proactive Detection and Prevention
By integrating this AI via an Application Programming Interface (API), financial institutions can perform realtime
analysis during the new account application process. AI provides an immediate risk score, allowing banks to predict whether a new account is likely to be used for fraud, stopping criminal activity before it starts.
High-Efficiency Performance
Following extensive training, the AI has achieved an accuracy rate of 85%. This means for every 100 attempted
mule account setups, the system successfully intercepts 85. This provides an innovative and highly effective
solution for blocking fraud at the source.
