AI-DRIVEN SIMULATION OF CYBER ATTACKS AND DEFENSES: A GENERATIVE APPROACH TO THREAT MODELING

Abstract

Cyber risk awareness is essential in today’s evolving threat landscape, where traditional detection methods like signature- and rule-based systems fall short against advanced, emerging attacks due to their reactive nature. To stay ahead, organizations must adopt proactive defense strategies. This paper explores the use of Generative AI—specifically models like GANs and VAEs—to simulate realistic cyberattacks and defense scenarios. These models can generate novel, previously unseen attack patterns, aiding in the assessment and enhancement of cybersecurity systems. By simulating threats such as APTs across the attack lifecycle, generative AI enables a shift from reactive to proactive threat modeling, aligning security capabilities with the growing sophistication of cyber threats.Generative AI