Leveraging AI to address SIM swap fraud during 2G/3G shutdown

Abstract

This research explores the potential of artificial intelligence (AI) as a safeguard against SIM (subscriber identity module) swap fraud in the evolving landscape of the global 2G/3G network shutdown. As the telecommunications industry transitions to more advanced 4G and 5G networks, phasing out of older 2G/3G infrastructure could inadvertently create new vulnerabilities that fraudsters may exploit. SIM swap fraud, where attackers gain unauthorised control of a victim’s mobile number, is particularly concerning during this period, as it enables interception of sensitive communications like one-time passwords, which can compromise financial accounts and personal information. Thus, this study assessed AI’s capacity to address these emerging risks. Using a systematic literature review grounded in the PRISMA framework, this research offers a comprehensive evaluation of the current knowledge base. Through bibliometric analysis, this work identifies key trends, influential publications, and leading researchers focused on AI-driven fraud prevention in telecommunications. This analytical approach yields a data-driven perspective on the field, highlighting influential areas and collaborative networks that shape the ongoing discourse around fraud mitigation amid network transitions. The findings indicate notable gaps in the existing literature and highlights high-impact studies, establishing a solid foundation for further research. In addition, a thematic analysis delves into recurring patterns across the reviewed literature, identifying significant themes such as AI’s role in anomaly detection, the prevention of social engineering attacks, and predictive analytics. This method categorises new fraud tactics arising during the 2G/3G shutdown, along with the specific security vulnerabilities that emerge in this transition phase. By examining these themes, the study illustrates AI’s potential to identify fraudulent activities in real time and strengthen security within customer service contexts, where social engineering remains a critical threat. v Ultimately, this research contributes actionable insights for AI-driven security measures in the telecommunications sector. The findings aim to inform telecom operators, financial institutions, and policymakers in crafting effective strategies to combat fraud, acknowledging the complex challenges that arise from decommissioning legacy networks.