AGENTIC AI AND CYBER SECURITY: AUTONOMOUS THREAT HUNTING, INTRUSION DETECTION, AND ADAPTIVE DEFENSE MECHANISMS IN A WORLD OF INCREASINGLY SOPHISTICATED CYBER ATTACKS

Abstract

This study investigates the transformative potential of agentic artificial intelligence (AI) systems in enhancing cybersecurity through autonomous threat hunting, real-time intrusion detection, and adaptive defense mechanisms. Employing a mixed-methods research design, the investigation analyzes a large-scale dataset comprising 2.4 million network events collected from January 2023 to December 2024 across 150 enterprise environments. A custom agentic AI framework built on reinforcement learning (RL), large language models (LLMs), and multi-agent collaboration was developed and evaluated against baseline machine learning models. Results demonstrate a 41% improvement in threat detection accuracy, a 57% reduction in mean time to respond (MTTR), and a 63% increase in adaptive policy efficacy under simulated advanced persistent threat (APT) conditions. The findings underscore the necessity of goal-directed, self-improving AI agents in countering evolving cyber threats while highlighting ethical, interpretability, and integration challenges. This work contributes a reproducible methodology and empirical benchmarks for deploying agentic AI in operational security environments.