Artificial Intelligence in Zero Trust Architecture for 5G and 6G Networks: a systematic literature review

Abstract

The transition to 5G and 6G networks has introduced a hyper-connected, virtualized environment where traditional perimeter-based security models were rendered obsolete by the dissolution of physical boundaries. The objective was to evaluate the necessity of Artificial Intelligence (AI) in operationalizing Zero Trust Architecture (ZTA) within these next-generation networks. A Systematic Literature Review (SLR) was implemented to synthesize findings from thirty-five primary studies, focusing on how AI functions as a mandatory enabler for dynamic policy enforcement rather than a simple enhancement. It was found that static rule-based systems failed to handle the velocity of 6G data, whereas AI-driven trust algorithms successfully reduced false positive rates and enabled necessary behavioural biometrics. However, a critical trade-off was identified between the computational overhead of these security models and the ultra-low latency requirements of the network. It was concluded that the "never trust, always verify" principle is effectively impossible to implement at scale without AI automation, establishing it as a foundational requirement for future telecommunications security despite the ongoing challenges regarding scalability and adversarial robustness.