Generative Artificial Intelligence for Smart IoT Systems: A Comprehensive Review

Abstract

The Internet of Things (IoT) has transformed conventional devices and environments into interconnected smart systems capable of sensing, communicating, and processing real-time data. However, the increasing scale, heterogeneity, resource constraints, security concerns, and massive volume of IoT-generated data create significant challenges for intelligent decision-making and autonomous operation. Generative Artificial Intelligence (GenAI) provides emerging capabilities for generating synthetic data, interpreting complex information, supporting prediction, enabling personalization, and facilitating human-machine interaction. This review presents a comprehensive analysis of Generative AI for Smart IoT Systems, covering IoT networking technologies, data-driven approaches, architectures, components, applications, and existing challenges. It further examines GenAI-enabled IoT architectures, including edge-cloud, federated, and autonomous approaches, as well as opportunities in model compression, runtime optimization, and edge-cloud collaboration. Applications in smart homes, healthcare, industrial IoT, smart cities, mobile networks, cybersecurity, and digital twins are discussed. Recent studies are comparatively analyzed to identify key findings, limitations, and research directions. Finally, the review highlights emerging opportunities and challenges for developing secure, scalable, efficient, adaptive, and intelligent GenAI-driven IoT ecosystems.