Securing Autonomous Vehicles with Smart AI Security Agents
Haidar Jabbar, Samir Al-Janabi, Francis Syms · 2025
As Autonomous Vehicles (AVs) become increasingly integrated into modern transportation systems, they face rising exposure to complex cybersecurity threats. Existing security frameworks, based on static rules, signature detection, and conventional Artificial Intelligence (AI), are insufficient in addressing the dynamic and adaptive nature of cyberattacks in AVs environment. This study presents a multi-layered defense approach powered by autonomous AI agents capable of independently detecting, interpreting, and responding to a wide range of cyber threats in real time. These agents employ Machine Learning (ML), Deep Learning (DL), anomaly detection, and adversarial resilience techniques to counter sophisticated attacks such as sensor manipulation, AI corruption, GPS spoofing, and Vehicle-to-Everything (V2X) based intrusions. The proposed architecture contributes to the advancement of intelligent, self-adjusting security mechanisms tailored for the evolving cybersecurity landscape of autonomous systems.