Automated Vulnerability Discovery Generative AI in Offensive Security
Mohammad Alauthman, Ammar Almomani, Samer Aoudi, Ahmad Al–Qerem, Amjad Yousef Aldweesh · Advances in computational intelligence and robotics book series · 2025
This chapter investigates how generative AI techniques transform offensive security. It explores automated methods that identify software flaws, generate targeted exploits, and support penetration testers in analyzing systems. Emphasis is placed on large language models and generative adversarial networks, which accelerate fuzzing processes and code reviews, uncovering complex or obscure vulnerabilities. Case studies detail how these AI-driven approaches drastically reduce time and effort compared to traditional methods, while also highlighting ethical concerns related to model hallucinations, dual-use threats, and privacy. The chapter concludes by examining emerging trends such as specialized security language models, fully autonomous red teams, and anticipated regulatory developments that could shape how generative AI evolves in this rapidly changing domain.