Abstract:

Smart cities are rapidly adopting advanced technologies to enhance life in cities, and cybersecurity plays a central role in safeguarding data and infrastructure. The application of Artificial Intelligence (AI) and Federated Learning (FL) can maximize security in smart cities by allowing decentralized, privacy-respecting solutions to detect and respond to cyber threats. Contemporary cybersecurity solutions are likely to be based on centralized models that gather massive amounts of data from a multitude of IoT devices, which increases the privacy, data integrity, and scalability risks. In addition, the increased intensity and complexity of cyberattacks mean that there is a need for an adaptive and real-time solution, which most of the existing systems cannot provide adequately. In this paper, a novel framework named Federated AI for Cybersecurity (FACS) is proposed, which integrates AI and FL to build a decentralized security platform for smart cities. The method ensures local processing of data on devices so that the threat of data breaches remains low and privacy is enhanced. FACS will enable real-time anomaly detection on diverse IoT devices with user privacy preserved. The findings show that the proposed approach fortifies cybersecurity through enhanced threat detection accuracy, risk reduction associated with centralized data storage, and effective response mechanisms. In addition, FACS is scalable and flexible in the ever-changing context of smart cities, providing reliable protection against diverse cyber threats.