Citation :

Vijay Kumar Soni, Aman Sardana, Pradeep Manivannan, 2025. "Context-Aware LLM Fraud Sentinels for Card Authorization" ESP International Journal of Advancements in Computational Technology (ESP-IJACT)  Volume 3, Issue 2: 36-49.

Abstract :

The rapid evolution of payment card fraud techniques necessitates advanced detection frameworks capable of adapting to emerging threats with minimal delay. Most of these systems have problems detecting unknown fraud patterns and allow much inaccurate positive detection. New hybrid architecture has been proposed using Large Language Models (LLMs), Graph Neural Networks (GNNs) and context together, which helps to detect fraud for the card authorization process promptly and reliably. The LLM section can extract meaningful metadata through using advanced methods, while the GNN component dynamically models transactional relationships and propagates risk scores across entities such as customers, merchants, and transactions. Because of this design, the method adapts better, achieves a balanced precision and recall and allows the system to understand and explain its actions through graph features and attention. When tested using anonymized transaction samples, the system achieved much better results in zero-day fraud, fewer false positives and shorter processing times. The research mentions that there are problems with scaling, privacy, ongoing training and explaining models and outlines how more research can focus on working together, using multiple data sources and testing regulations in practice. This integrated approach lays the foundation for next-generation, context-aware fraud detection systems that can safeguard payment ecosystems while delivering seamless user experiences.

References :

[1] C. Scardovi, Digital Transformation in Financial Services, vol. 236. Cham: Springer International Publishing, 2017.

[2] Sumanjeet, "Emergence of payment systems in the age of electronic commerce: The state of art," in 2009 First Asian Himalayas International Conference on Internet, IEEE, 2009.

[3] R. F. Olanrewaju, et al., "Securing electronic transactions via payment gateways–a systematic review," Int. J. Internet Technol. Secur. Transact., vol. 7, no. 3, pp. 245–269, 2017002E

[4] K. G. Al-Hashedi and P. Magalingam, "Financial fraud detection applying data mining techniques: A comprehensive review from 2009 to 2019," Comput. Sci. Rev., vol. 40, p. 100402, 2021.

[5] R. Khurana, "Fraud detection in ecommerce payment systems: The role of predictive AI in real-time transaction security and risk management," Int. J. Appl. Mach. Learn. Comput. Intell., vol. 10, no. 6, pp. 1–32, 2020.

[6] A. Papasavva, et al., "Application of AI-based Models for Online Fraud Detection and Analysis," arXiv preprint arXiv:2409.19022, 2024.

[7] M. Fan, "LLMs in Banking: Applications, Challenges, and Approaches," in Proc. Int. Conf. Digit. Econ., Blockchain Artif. Intell., 2024.

[8] B. Yadav, "Generative AI in the Era of Transformers: Revolutionizing Natural Language Processing with LLMs," J. Image Process. Intell. Remote Sens., vol. 4, no. 2, pp. 54–61, 2024.

[9] S. Borgeaud, et al., "Improving language models by retrieving from trillions of tokens," in Proc. Int. Conf. Mach. Learn. (ICML), PMLR, 2022.

[10] N. Wang, et al., "Deep compression of pre-trained transformer models," Adv. Neural Inf. Process. Syst., vol. 35, pp. 14140–14154, 2022.

[11] A. M. Agrawal, "Transforming e-commerce with Graph Neural Networks: Enhancing personalization, security, and business growth," in Applied Graph Data Science. Morgan Kaufmann, 2025, pp. 215–224.

[12] H. Matsumoto, S. Yoshida, and M. Muneyasu, "Propagation-based fake news detection using graph neural networks with transformer," in 2021 IEEE 10th Global Conf. Consum. Electron. (GCCE), IEEE, 2021.

[13] Z. Song, Y. Zhang, and I. King, "Towards fair financial services for all: A temporal GNN approach for individual fairness on transaction networks," in Proc. 32nd ACM Int. Conf. Inf. Knowl. Manag., 2023.

[14] U. Rajeshwari and B. S. Babu, "Real-time credit card fraud detection using streaming analytics," in 2016 2nd Int. Conf. Appl. Theor. Comput. Commun. Technol. (iCATccT), IEEE, 2016.

[15] E. A. Morse and V. Raval, "PCI DSS: Payment card industry data security standards in context," Comput. Law Secur. Rev., vol. 24, no. 6, pp. 540–554, 2008.

[16] B. Vagadia, "Data integrity, control and tokenization," in Digital Disruption: Implications and Opportunities for Economies, Society, Policy Makers and Business Leaders, Cham: Springer Int. Publ., 2020, pp. 107–176.

[17] G. Yenduri, et al., "GPT (Generative Pre-trained Transformer)–A comprehensive review on enabling technologies, potential applications, emerging challenges, and future directions," IEEE Access, 2024.

[18] A. Khazane, et al., "Deeptrax: Embedding graphs of financial transactions," in 2019 18th IEEE Int. Conf. Mach. Learn. Appl. (ICMLA), IEEE, 2019.

[19] X. Mao, M. Liu, and Y. Wang, "Using GNN to detect financial fraud based on the related party transactions network," Procedia Comput. Sci., vol. 214, pp. 351–358, 2022.

[20] S. Wang, et al., "Optimizing logical execution time model for both determinism and low latency," in 2024 IEEE 30th Real-Time Embedded Technol. Appl. Symp. (RTAS), IEEE, 2024

[21] C. Chen, et al., "A survey on graph neural networks and graph transformers in computer vision: A task-oriented perspective," IEEE Trans. Pattern Anal. Mach. Intell., 2024.

[22] G. Baader and H. Krcmar, "Reducing false positives in fraud detection: Combining the red flag approach with process mining," Int. J. Account. Inf. Syst., vol. 31, pp. 1–16, 2018.

[23] S. Wang, et al., "Graph machine learning in the era of large language models (LLMs)," ACM Trans. Intell. Syst. Technol., 2024.

[24] A. El Bouchti, et al., "Fraud detection in banking using deep reinforcement learning," in 2017 Seventh Int. Conf. Innov. Comput. Technol. (INTECH), IEEE, 2017.

[25] Z. Liu, et al., "Privacy-preserving aggregation in federated learning: A survey," IEEE Trans. Big Data, 2022.

Keywords :

Fraud Detection, Large Language Models (LLMs), Graph Neural Networks (GNNs), Zero-Day Fraud, Explainable AI (XAI).