An Adaptive Cybersecurity Framework Using Generative Artificial Intelligence for Intelligent Threat Detection

DOI

10.2991/978-94-6239-789-7_26How to use a DOI?

Keywords

Cybersecurity; Generative Artificial Intelligence; Intrusion Detection System; Zero-Day Attacks; Autonomous Response; Real-Time Security

Abstract

The evolution of cyber threats like zero-day attacks and polymorphic malware has made it increasingly challenging for traditional intrusion detection systems, which depend on static signatures and supervised learning methods, to keep pace. These traditional systems are limited in their adaptability, suffer from data imbalance issues, and are unresponsive to real-time situations. This research provides a new Adaptive Cybersecurity Framework powered by Generative Artificial Intelligence (GenAI), allowing for intelligent threat detection and an autonomous response to these threats. We have created an Adaptive Cybersecurity Framework including live network traffic capture, Generative Adversarial Networks (GAN)-based synthetic attack generation, adaptive Machine Learning (ML)-based Intrusion Detection, automated response to detected threats and the ability for continual adaptation of the Intrusion Detection System (IDS) utilizing detected events as input to the learning process. In our model, the Generative Module creates realistic, never-before-seen attack patterns based on the characteristics of generated attack traffic, allowing for greater diversity in dataset representations, as well as improving the ability to identify threats that are rare or are zero-days. The Hybrid CNN/LSTM Detection Engine learns the spatial-temporal characteristics of network traffic by continually adapting based on the data collected from previous events. We have demonstrated the ability to implement and validate the framework with data from multiple datasets including CICIDS 2017, UNSW-NB15, and NSL-KDD, through the use of real-time network traffic. The experimental findings show that as compared to baseline deep-learning IDS’s, the Adaptive Cybersecurity Framework demonstrated an increase in accuracy of 15.2%, an increase in F1-Score of 18.7%, and a decrease in Response Latency of 27.4%. Overall, the results of this research confirm that the Adaptive Cybersecurity Framework presents an effective, robust and applicable model for Cybersecurity in modern network environments.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Rohan Chaudhary
AU  - Ankur Rana
AU  - Himanshu Tyagi
PY  - 2026
DA  - 2026/10/08
TI  - An Adaptive Cybersecurity Framework Using Generative Artificial Intelligence for Intelligent Threat Detection
BT  - Proceedings of theInternational Conference on Innovation and Practices in Artificial Intelligence (ICIPAI 2026)
PB  - Atlantis Press
SP  - 380
EP  - 393
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-789-7_26
DO  - 10.2991/978-94-6239-789-7_26
ID  - Chaudhary2026
ER  -