Context-Aware AI Models for Anomaly Detection in Encrypted 5G and IoT Networks

The rapid deployment of 5G and Internet of Things (IoT) technologies has transformed digital connectivity but also introduced complex cybersecurity challenges. As data transmission increasingly occurs over encrypted channels, conventional intrusion detection methods that rely on payload inspection have become less effective. This study presents a context-aware artificial intelligence (AI) framework designed to detect anomalies in encrypted 5G-IoT traffic without violating user privacy. The proposed model integrates flow-based statistical features with contextual metadata such as device behavior, communication frequency, and temporal interaction patterns to capture deviations from legitimate network behavior. Machine learning and deep learning models are trained using real and simulated datasets, with explainability components incorporated to interpret the decision process and support transparency in threat assessment. Experimental evaluation demonstrates that context-aware modeling significantly improves detection accuracy and reduces false positives compared to baseline encryption-agnostic systems. The findings highlight the potential of context-driven AI analytics to enhance secure, privacy-preserving monitoring across next-generation 5G and IoT infrastructures.