Detecting Cyber Threat Using Gen-AI
T.Nithya Shree, K. Sundarakantham, J. Josephine Dhivya, B. Gayathri · 2025
As we stand in the era of 6G, the ultra-fast and low-latency connectivity opens up numerous possibilities for innovation, across various sectors. These advancements bring more sophisticated cyber threats. The complexity and scale of 6G networks demand innovative approaches to cybersecurity, and this is where generative AI (GenAI)can play a transformative role. Traditional cybersecurity measures such as signature-based detection and heuristic analysis, have served us well in the past. The deep learning models like BERT have numerous in cyber threat detection, the need for efficient detection methods is critical because of more frequent and complex threats, With differential privacy with embedding perturbation technique(DPEP). The Bert model provides a cutting-edge architecture to detect threats with high precision. Through the experiments held on the Edge-IIoT set cybersecurity dataset, Bert has achieved 91.2% accuracy. Thus, Bert might be an optimal solution for real-time threat detection, especially for deploying resource-constrained loT environments.