Graph Neural Network-Based Detection for Anti-Money Laundering in Financial Transaction Networks: A Systematic Review for Continual Learning and Explainability in Resource-Constrained Regulatory Environments
Keywords:
anti-money laundering; graph neural networks; continual learning; explainable artificial intelligence; PRISMA 2020Abstract
Anti-money laundering (AML) detection has evolved from static rule-based engines to Graph Neural Network (GNN) architectures able to model the relational structure of financial transactions. This systematic review, conducted under the PRISMA 2020 statement (Page et al., 2021), synthesises 26 years (2000–2026) of graph-based and network-analytic AML research to trace this evolution and establish whether continual learning and explainability have been jointly integrated and validated within GNN-based AML systems. Five databases (Scopus, Web of Science, IEEE Xplore, ACM Digital Library and arXiv) were searched using three Boolean strings covering graph/network methods, continual learning and explainability, supplemented by citation snowballing. Of 81 records identified, 46 remained after deduplication and 24 met full-text eligibility, screened against four PRISMA criteria and appraised with a six-item modified Joanna Briggs Institute checklist (Joanna Briggs Institute, 2020). Studies were synthesised narratively across five evolutionary generations: rule-based systems, graph-theoretic/network analysis, hybrid graph-machine-learning, GNN-based detection, and integrated continual-learning/explainability frameworks. Detection performance improved markedly, from false-positive rates of 90–95% in rule-based systems (OECD, 2009) to AUC-ROC scores of 0.874–0.945 in recent attention-based and Transformer-augmented GNNs (Olaniyi et al., 2026). However, no study integrates relational detection, continual learning and regulatory-grade explainability simultaneously. Three gaps persist across all generations: fragmented relational-temporal modelling, an unresolved tension between static and adaptive deployment, and a reproducibility deficit caused by reliance on proprietary data. The review recommends that future work jointly evaluate experience-replay continual learning (Deprez et al., 2025) with GNNExplainer/SHAP explainability (Ying et al., 2019) on open, reproducible benchmarks, and test explanation stability under incremental model updates and severe class imbalance, so as to close the gap between regulatory expectations and current GNN-AML capability.
References
Ajagbe, S. A., Majola, S., & Mudali, P. (2025). Comparative analysis of machine learning algorithms for money laundering detection. Discover Artificial Intelligence, 5(1), 144.
Bakhshinejad, N., Nguyen, U., Ghahremani, S., & Soltani, R. (2024). A graph-based deep learning model for the anti-money laundering task of transaction monitoring. Proceedings of the 16th International Joint Conference on Computational Intelligence, 496–507.
Cardoso, M., Saleiro, P., & Bizarro, P. (2022). LaundroGraph: Self-supervised graph representation learning for anti-money laundering (arXiv:2210.14360). arXiv.
Cassara, J. A. (2015). Frontmatter. In Trade-based money laundering (pp. i–xx). John Wiley & Sons.
Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20(1), 37–46.
Deprez, B., Vanderschueren, T., Baesens, B., Verdonck, T., & Verbeke, W. (2025). Network analytics for anti-money laundering—A systematic literature review and experimental evaluation. INFORMS Journal on Data Science.
Dreżewski, R., Sepielak, J., & Filipkowski, W. (2012). System supporting money laundering detection. Digital Investigation, 9(1), 8–21.
FinCEN. (n.d.). What is money laundering? Retrieved January 17, 2026, from https://www.fincen.gov/what-money-laundering
Hamilton, W. L., Ying, R., & Leskovec, J. (n.d.). Inductive representation learning on large graphs.
He, X., Huang, J., Ma, K., He, H., & Li, M. (2026). An explainable graph neural network framework for illicit financial transaction detection. Applied Intelligence, 56(4), 102.
Joanna Briggs Institute. (2020). Critical appraisal tools. JBI. https://jbi.global/critical-appraisal-tools
Kipf, T. N., & Welling, M. (2017). Semi-supervised classification with graph convolutional networks.
Krebs, V. E. (2002). Mapping networks of terrorist cells.
Lawal, O., Okolie, A., & Obunadike, C. (2025, December 17). An explainable graph neural network framework for anti–money laundering in cryptocurrency transactions using the Elliptic dataset. International Journal of Computer Networks & Communications.
Levi, M., Ang, D., Gundur, R., & Williams, M. (2015). The implications of economic cybercrime for policing.
Levi, M., & Reuter, P. (2006). Money laundering. Crime and Justice, 34, 289–375.
Lopez-Rojas, E. A., & Axelsson, S. (n.d.). Money laundering detection using synthetic data.
Marasi, S., & Ferretti, S. (2024). Anti-money laundering in cryptocurrencies through graph neural networks: A comparative study. In Proceedings of the 2024 IEEE 21st Consumer Communications & Networking Conference (CCNC). IEEE.
OECD. (2009). Money laundering awareness handbook for tax examiners and tax auditors. OECD.
Olaniyi, O. M., Aroh, I. S., Henry, O., Metibemu, O. C., & Akinola, O. I. (2026). Graph neural networks for multi-layered financial crime network detection: An explainable AI framework for anti-money laundering. Journal of Engineering Research and Reports, 28(2), 18–36.
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., … Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, n71.
Savage, D., Zhang, X., Wang, Q., Yu, X., & Chou, P. (2016). Detection of money laundering groups: Supervised learning on small networks.
Sharafaldin, I., Habibi Lashkari, A., & Ghorbani, A. A. (2018). Toward generating a new intrusion detection dataset and intrusion traffic characterization. Proceedings of the 4th International Conference on Information Systems Security and Privacy, 108–116.
UNODC. (2011). Estimating illicit financial flows resulting from drug trafficking and other transnational organized crimes. Office of Justice Programs.
Velickovic, P., Cucurull, G., Casanova, A., Romero, A., Lio, P., & Bengio, Y. (2018). Graph attention networks.
Weber, M., Domeniconi, G., Chen, J., Weidele, D. K. I., Bellei, C., Robinson, T., & Leiserson, C. E. (2019). Anti-money laundering in Bitcoin: Experimenting with graph convolutional networks for financial forensics.
Ying, R., Bourgeois, D., You, J., Zitnik, M., & Leskovec, J. (2019). GNNExplainer: Generating explanations for graph neural networks (arXiv:1903.03894). arXiv.
Yu, Q., Wang, S., & Tao, Y. (2025). Enhancing anti-money laundering detection with self-attention graph neural networks. SHS Web of Conferences, 213, 01016.
Zaslavsky, V., & Strizhak, A. (2006). Credit card fraud detection using self-organizing maps. Information & Security: An International Journal, 18, 48–63.
Zhang, W., Shen, Y., Zhou, H., Zhou, B., Zheng, X., & Chen, X. (2026). Temporal attentive graph networks for financial surveillance: An incremental multi-scale framework. Journal of Sensor and Actuator Networks, 15(1), 23.
