Artificial Intelligence in Cyber Defense: A Novel Honeypot Architecture for Real-Time Threat Engagement Using AI Agents In A Hierarchy Structure

Modern cyber threats continue to evolve, making traditional honeypots less effective as attackers adapt and bypass static systems. This paper introduces a novel honeypot framework powered by artificial intelligence, designed to engage threats in real time and adapt to attacker behavior. Unlike conventional honeypots that follow pre-scripted patterns, this system uses machine learning to analyze and mimic target environments dynamically, increasing realism and prolonging attacker interaction. The architecture includes automated decision-making, behavioral deception, and modular threat response, providing both detection and data collection capabilities. Preliminary design and testing indicate that the AI honeypot is capable of adjusting to novel attack methods, enhancing both threat visibility and response effectiveness. This approach aims to close the gap between static traps and dynamic attackers, contributing to the next generation of cybersecurity defense tools.