AI Guards EV Chargers From Cyber Threats

The rapid expansion of electric vehicle (EV) charging infrastructure is introducing significant cybersecurity vulnerabilities that remain largely unaddressed. These complex systems, integrating multiple physical and digital elements, are exposed to a range of attacks, threatening both EV adoption and national electrical grid stability.

To combat these evolving threats, researchers at the University of Malaga have proposed an innovative solution employing AI agents. These agents are designed to detect and prevent cyberattacks, ranging from localized fraud and energy theft to larger-scale assaults on critical energy networks. The system leverages the Open Charge Point Protocol (OCPP) to identify anomalies and attacks within charging networks.

Each component of the charging network is equipped with AI agents that analyze their environment, gather data, and collaborate with other agents. This collective intelligence builds a comprehensive understanding of the infrastructure's status, allowing for more accurate detection of operational failures and security incidents. A novel consensus mechanism, based on opinion dynamics, helps agents reach collective agreements and reduces the likelihood of false positives.

Further enhancing security, blockchain technology is used as a trust and validation mechanism, ensuring the integrity and traceability of all agent transactions through an immutable ledger. A simulated testing environment demonstrated that this integrated system effectively detects specific anomalies and broader behavioral patterns across multiple charging stations, improving diagnostic accuracy and network reliability.