AI-Assisted Cybersecurity Mesh for Threat Detection in Edge-Enabled Communication Networks

Abstract

Next-generation communication environments increasingly combine IoT devices, edge gateways, cyber-physical components, and programmable network services. This convergence improves responsiveness but also creates fragmented trust boundaries and fast-changing attack surfaces. Conventional intrusion detection systems remain limited in such settings because they depend heavily on static signatures, centralized telemetry collection, or offline machine-learning models. This paper introduces a GenAI-assisted cybersecurity mesh for threat detection in heterogeneous intelligent communication systems. The framework places lightweight security functions at edge nodes, coordinates them through a mesh control layer, and uses a generative threat modeling engine to update anomaly assumptions as traffic conditions change. Network, application, and behavioral signals are fused into a dynamic risk score that supports policy actions such as throttling, isolation, and micro-segmentation. The framework is evaluated in a simulated communication environment with mixed benign traffic and attack scenarios, including DDoS, man-in-the-middle, protocol exploitation, behavioral drift, and synthetic zero-day patterns. Results show higher detection accuracy, lower false positive rates, and reduced response latency compared with rule-based and centralized ML-based IDS baselines. The study positions cybersecurity mesh as a practical direction for low-latency, AI-assisted protection of distributed communication infrastructures.