The rapid proliferation of Internet of Things (IoT) devices across home, office, and network infrastructure ecosystems has introduced significant security challenges, particularly regarding the presence of rogue or unauthorized devices. Due to the dynamic nature of modern networks, traditional security techniques frequently fail to detect these unauthorized assets. To address this, behavioral fingerprinting offers an effective solution for device identification by analyzing network activity, leveraging the fact that different devices exhibit unique communication patterns. This review paper focuses on Generative AI based approaches for behavioral fingerprinting in IoT networks. Specifically, we examine existing literature on anomaly detection and device identification that utilizes methodologies such as Autoencoders, Variational Autoencoders (VAEs), and Generative Adversarial Networks (GANs). Furthermore, this paper analyzes the performance metrics, computational resource requirements, and security concerns associated with implementing Generative AI within IoT environments.