CYBER THREAT FORECASTING :THE TRANSITION FROM TRADITIONAL ML TO GENERATIVE AL

Abstract

Cyber threat forecasting has emerged as a crucial component of modern cybersecurity strategies due to the increasing complexity, frequency, and sophistication of cyber attacks. Traditional machine learning (ML) techniques have been widely used for threat detection, anomaly identification, and risk prediction by analyzing historical security data and network behavior patterns. However, the rapidly evolving nature of cyber threats requires more advanced and adaptive forecasting mechanisms capable of understanding complex attack scenarios and generating predictive insights. This paper explores the transition from traditional machine learning to Generative Artificial Intelligence (Generative AI) for cyber threat forecasting. The proposed framework leverages the predictive capabilities of conventional ML models alongside the advanced reasoning, simulation, and content-generation abilities of Generative AI to forecast emerging cyber threats, attack trends, and potential vulnerabilities. Generative AI enables the creation of realistic threat scenarios, automated threat intelligence analysis, and proactive security recommendations. Experimental observations indicate that Generative AI enhances forecasting accuracy, improves threat anticipation capabilities, and supports more effective cybersecurity decision-making. The integration of traditional ML and Generative AI provides a powerful approach for developing intelligent, adaptive, and future-ready cyber defense systems.