Graph neural network improves money laundering detection with clear explanations
Explainable and Trustworthy AI for Anti-Money Laundering: A Graph-based Hybrid Framework for Real-World Financial Crime Detection
Computational Engineering, Finance, and Science
Summary
Money laundering harms financial systems but is hard to detect because usual rules raise many false alarms and miss complex transaction patterns. The authors built a graph-based AI model that looks at relationships between transactions to better spot suspicious activity. Their best model was much more accurate than others, explained why it made decisions faithfully, and assessed how confident it was about its choices. This helps people in finance better understand risks and make informed decisions about potential money laundering.
What this means in practice
- •For financial compliance teams: Use graph neural networks to improve detection accuracy and transparency of suspicious transactions for regulatory compliance.
- •For fraud detection engineers: Incorporate topology-aware AI models with explanation tools to reduce false positives and support human review in financial crime detection.
Authors
Ali Shahbazi, Faraz Sasani, Arshia Hossein zadeh, Sheyda safaeimoradi, Hossein Najafzadeh
Abstract
Purpose: Money laundering threatens financial systems, while rule-based monitoring suffers from high false-positive rates and limited ability to capture relational transaction patterns. This study proposes an explainable graph-based framework for anti-money laundering (AML) detection that jointly addresses predictive performance, explanation faithfulness, and uncertainty calibration. Methods: Using the IBM Transactions for Anti-Money Laundering (HI-Small) benchmark, comprising 5,078,345 transactions among 518,573 accounts with 0.10% illicit transactions, a directed attributed graph with 4,487,133 edges was constructed using temporal, leakage-safe partitioning. Three GATv2 architectures, BASE, BASE-Large, and IMPROVED, incorporating bidirectional message passing, edge updates, and port-aware features, were compared with XGBoost and Random Forest using identical features. The best model was evaluated using GNNExplainer against documented typologies and Mondrian conformal prediction for uncertainty calibration. Results: IMPROVED achieved the strongest performance (AUPRC = 64.85%, Best F1 = 68.42%), exceeding BASE-Large by 30.65 AUPRC points and XGBoost (AUPRC = 38.46%) by 26.4 points. The proposed mechanisms contributed more than capacity scaling alone. GNNExplainer recovered documented typologies with higher fidelity than attention-weight and random baselines (mean Jaccard overlap: 21.4% vs. 1.0%, p < 0.001). Mondrian conformal prediction achieved coverage close to the 90% nominal target with an average prediction-set size near one. Conclusion: Explicit transaction topology modeling substantially improves AML detection, while faithful explanations and calibrated uncertainty support interpretable, human-in-the-loop compliance decision-making.