AUTOMATED CYBER THREAT INTELLIGENCE ANALYSIS AND DETECTION USING INTEGRATED GENERATIVE AI AND RETRIEVAL-AUGMENTED FRAMEWORK

Abstract

Due to increase in the spread of cyber threats, there is a requirement of creating an intelligent cyber threat mechanism to combat threats. Existing methods are typically rule-based and manual, which makes them insufficient for detecting new threats. This research aims to create a Generative AI-based framework for automated threat analysis and detection using the combination of Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG). The framework receives cybersecurity data from various sources, identifies IoCs, fetches threat intelligence and classifies the threats in this context. The framework also produces enriched threat intelligence reports which include identified threats, attack techniques, supporting evidence and defensive recommendations. Its effectiveness will be measured through the accuracy, precision, recall, F1-score, hallucination rate, evidence relevance, and processing time.