Artificial Intelligence Based Anomaly Detection for Secure E-Government Transaction: A Review

Abstract

The high pace of e-government platforms has revolutionized the way of delivering public services because it has facilitated massive online transactions in authentication, finance, welfare disbursement, land management, and citizen-government interfaces. The growing reliance on digital governance has however demonstrated vulnerability of systems to advanced cyber threats such as data manipulation, identity fraud, access without authorization, malware infiltration and huge-scale coordinated attacks that are beyond the ability of traditional rule based security mechanisms to manage. The present review article explores why the use of Artificial Intelligence (AI)-based anomaly detection is critical in increasing the security, reliability, and resilience of e-government transactions. Through examination of advanced machine learning and deep learning frameworks, including LSTM, autoencoders, graph neural networks, and unsupervised clustering models, the researcher can point out that AI systems are able to learn a behavioral pattern, identify anomalies in real-time, and detect threats pretending to be unknown or zero-day threats with high levels of precision. The literature also indicates that there are new frameworks in the emerging of combining blockchain and artificial immune system, federated learning and hybrid cloud and edge architecture to enhance data integrity and privacy. The results indicate that AI-based anomaly detection offers a scalable, adaptive, and proactive protection mechanism that is necessary in the protection of the modern e-government ecosystems. The paper ends by arguing why ongoing innovation and cross-functional integration are necessary towards developing safe, reliable, and future intensive digital governance structures.