Collaborative Anti-Money Laundering Using Federated Learning

This conference talk explores how federated learning can revolutionize anti-money laundering (AML) efforts through cross-institutional collaboration. Presented by Kristiina Ausmees, Data Scientist at Handelsbanken, and Edvin Callisen, Research Scientist at AI Sweden, the presentation addresses the growing challenge of money laundering, estimated at $3.1 trillion globally in 2023. Learn about the limitations of traditional siloed, rule-based AML systems and discover a holistic framework that overcomes key challenges in data sharing, collaboration, and explainability. The speakers demonstrate a pipeline for creating synthetic datasets that simulate realistic inter-bank heterogeneity and show how federated learning improves detection performance while addressing practical implementation concerns for banks. Recorded at the 2025 GAIA Conference in Gothenburg, Sweden, this 25-minute talk offers valuable insights for financial institutions and regulatory professionals seeking to enhance their AML capabilities through advanced machine learning techniques.