NLP and Explainable AI in Credit Card Fraud Detection: Methods, Applications, and Real-Time Challenges

This systematic review establishes six key findings: (1) NLP enhances fraud detection accuracy by 5-15% through text-based features from transaction descriptions and customer communications [2], [3]; (2) XAI methods enable transparency—SHAP and LIME successfully explain fraud decisions, with attention-based models providing inherent interpretability [5]; (3) Combined NLP+XAI achieves 95% precision while maintaining interpretability [2], [9]; (4) Real-time deployment requires substantial optimization through model distillation, quantization, and selective application [2], [14]; (5) Graph Neural Networks represent an emerging frontier with 99% accuracy through transaction network structure capture [7], [8]; and (6) Regulatory compliance increasingly drives explainability as a requirement rather than optional enhancement [1].

[1]      FinCrime Central Editorial. (2025). Why AI explainability is becoming a regulatory imperative in AML and financial crime prevention. FinCrime Central, September 10, 2025. [2]      Longo, L., & Raufi, B. (2024). A comparative analysis of SHAP, LIME, ANCHORS, and DICE for interpreting a dense neural network in credit card fraud detection. Technological University Dublin Conference Proceedings, 1453, 1-18. [3]      Frontiers in Artificial Intelligence Editorial Board. (2025). Enhancing credit card fraud detection using traditional and deep learning approaches. Frontiers in Artificial Intelligence, 8, 1643292. [4]      Ranade, S., & Mehta, A. (2025). Comparative analysis of machine learning models for credit card fraud detection with SMOTE. International Journal of Sustainable Smart and Innovative Systems Engineering, 15(5), 504-518. [5]      Ojo, I. P., & Tomy, A. (2025). Explainable AI for credit card fraud detection: Bridging the gap between accuracy and interpretability. World Journal of Advanced Research and Reviews, 25(2), 1246-1256. [6]      Rajkumar, P., & Sinha, R. (2024). Optimizing credit card fraud detection with random forests and neural networks. PMC/NIH Research Publications, 12098799, 1-18. [7]      Xiang, Y., Li, W., & Cheng, D. (2024). Graph neural networks for financial fraud detection: A comprehensive review of methodologies and applications. arXiv preprint arXiv:2411.05815. [8]      Ding, X., Lv, H., Zhao, Y., Rao, R. M., & Zhu, Y. (2024). Real-time transaction fraud detection via heterogeneous temporal graph neural networks. Proceedings of International Conference on AI Finance, 365-378. [9]      Khandelwal, C. (2024). Multi-model stacking ensemble framework for financial fraud detection. International Journal of Innovative Research in Computer and Communication Engineering, 12(8), 2502-2520. [10]      Martini, R., & Barua, A. (2020). Deep recurrent neural networks for fraud detection on debit card transactions. Edinburgh Business School CRC Working Papers, 29, 1-15.