AI-Driven Threat Detection and Mitigation in Next-Generation Communication Networks
Kartheek Dokka, Ruchi Mangharamani, Bharath Thandalam Rajasekaran, Pranita Singh, Saravanan Thirumazhisai Prabhagaran, Arpit Kumar Jain · 2025
The introduction of next-generation communications networks, i.e., 5G and beyond, introduced unprecedented connectivity and scalability at the cost of a greater degree of complexity in securing these digital infrastructures. This work analyzes how Artificial Intelligence (AI) and particularly deep learning (DL) can assist in improving threat detection and mitigation for these systems. We propose an architecture that employs machine learning (ML) and deep learning (DL) algorithms to continuously monitor network activity, identify anomalies, and proactively respond to possible vulnerabilities. As is clear from our experimental results, Convolutional Neural Networks (CNNs) have the highest total accuracy (approximately 97%) and F1-score (around 96.8%), while Recurrent Neural Networks (RNNs) demonstrate very good recall (around 97%), and Long Short-Term Memory (LSTM) models are still competitive in terms of precision and exhibit balanced performance in terms of all metrics. The results demonstrate the potential of AI-based techniques for real-time threat detection and autonomous response. We then discuss how AI can be integrated with existing security systems, and the privacy as well as ethical implications. And finally, we discuss regulatory compliance and challenges in deploying a scalable, flexible, and authentic model validation-based AIbased security system in next generation networks.