HuntGPT: AI-Based Intrusion Detection Tool
Critical infrastructure has become increasingly the target of cyberattacks, with anticipated yearly damages of $10.5 trillion USD by 2025, up from only $3 trillion USD in 2015. NIST introduced a Cybersecurity Framework in 2014 to address these evolving threats.
Machine learning-based anomaly detection tools uncover both known and unknown threats, including performance and security anomalies. However, they often increase false positives in real-world use.
Large Language Models (LLMs) are poised to revolutionize cybersecurity by seamlessly integrating AI tasks and reducing operational costs. Their adaptability and role in actionable AI make them valuable for threat response.
Cybersecurity analysts Tarek Ali and Panos Kostakos from the Information Technology and Electrical Engineering Center for Ubiquitous Computing University of Oulu recently reported about HutGPT, an AI-based intrusion detection tool.
HuntGPT, a dashboard with a Random Forest classifier trained on KDD99, utilizes XAI frameworks like SHAP and Lime for enhanced user-friendliness. With GPT-3.5 Turbo, it presents detected threats in an easily explainable format.
Source: HuntGPT