Fraud Prevention
Status:
The project starts in October 2026.
Researchers:
Dr. Daniel Donatsch (ZKB)
Industry Partner:
ZKB
Description:
Fraud is a significant problem in the financial sector. Examples of fraud include scams where victims are tricked into making payments (e.g., romance scams, investment scams), as well as scenarios where criminals execute payments using stolen or hijacked credentials. Further examples include the creation of fake accounts and the malicious migration of accounts.
Most banks and financial institutions tackle fraud using a combination of rules-based detection systems and machine learning (ML) models, after which cases are analyzed by human case managers. However, the solutions used in the industry have limitations and fail to detect many complex cases of fraud. Moreover, case managers often need to handle many cases of potential fraud with limited support.
The goal of this project is to study fraud detection systems that include detection rules, ML components, and human case managers. The aim is to both improve the automatic detection of potential fraud cases through better ML models and rules and to develop additional technological support for case managers.