IDS-GPT: A Novel Deep Learning-Powered Framework for Network Traffic Intrusion Detection
Firas Saidi · Procedia Computer Science · 2025
Intrusion Detection Systems (IDS) play a pivotal role in network security, serving as essential tools to detect and mitigate malicious activities, thereby ensuring the protection of systems and the integrity of sensitive data. The rapid evolution of artificial intelligence, particularly in the realms of machine Learning (ML) and Deep Learning (DL), has ushered in groundbreaking methodologies that substantially improve the effectiveness of IDS. This paper investigates the utilization of ChatGPT, an advanced deep learning-driven language model, for the detection of malicious intrusions through the analysis of network traffic. By harnessing ChatGPT’s superior contextual understanding and analytical capabilities, we introduce IDS-GPT as an innovative model that achieves remarkable precision in identifying anomalies within network traffic. Rigorous testing on benchmark dataset CICIDS2018 and comparison to existing techniques such as Decision Tree, CNN-BiLSTM and GBM, reveal IDS-GPT ’s outstanding performance in detecting malicious activities, delivering high accuracy while maintaining exceptionally low false-positive rates. The findings, supported by a robust suite of precise evaluation metrics, underscore the transformative impact of integrating ChatGPT into IDS frameworks. This advancement sets the stage for the development of more resilient and intelligent cybersecurity solutions, capable of addressing the growing complexities of modern network threats.