Stefan Donsa & Lukas Alber - Detecting Money Laundering with AI

#1about 5 minutes

The lifecycle for operationalizing AI models in business

Moving beyond local development requires a structured lifecycle including ideation, proof of value, scaling, and continuous monitoring.

#2about 3 minutes

Understanding the limitations of rule-based AML systems

Traditional rule-based approaches for anti-money laundering suffer from high false-positive rates and fail to capture complex laundering patterns.

#3about 2 minutes

How AI improves AML and the challenges involved

AI can lower false positives and identify new threats, but success requires involving business experts and using explainable AI to build trust.

#4about 4 minutes

Using machine learning to detect KYC inconsistencies

Machine learning models analyze peer group behavior to identify outliers, such as a jobless person with high cash transactions, which rule-based systems miss.

#5about 5 minutes

A four-step process for unsupervised outlier detection

The process involves selecting relevant features, creating a master data table, using dimensionality reduction to find outliers, and scoring customers by reconstruction error.

#6about 3 minutes

Comparing PCA and autoencoders for anomaly detection

Principal Component Analysis (PCA) uses linear transformations while autoencoders use non-linear transformations for dimensionality reduction and reconstruction.

#7about 2 minutes

Validating the model's effectiveness with real-world results

The unsupervised models successfully identified over 100 suspicious cases, with one-third being new discoveries not caught by the existing rule-based engine.

#8about 3 minutes

The production architecture and technology stack for AML AI

The end-to-end system uses Hadoop for the data mart, PySpark for transformation, and Python with scikit-learn and MLflow for model development.

#9about 1 minute

Key takeaways for implementing AI in financial compliance

AI enhances AML efforts by detecting novel patterns, focusing agents on high-risk alerts, and providing transparent results to build trust with regulators.

today

Senior Agentic Data Scientist