Generative artificial intelligence represents a fundamental shift in medical image analysis by enabling the synthesis, reconstruction, and enhancement of imaging data rather than merely interpreting existing images. In oral and maxillofacial radiology, the adoption of generative models addresses critical challenges such as limited availability of annotated datasets, image degradation due to metal artifacts, and the need for predictive visualization of disease progression and treatment outcomes. This chapter presents a comprehensive overview of generative AI frameworks relevant to maxillofacial imaging, with a focus on generative adversarial networks (GANs), variational autoencoders (VAEs), and diffusion-based models. The theoretical foundations of these architectures are discussed alongside their clinical applications, including synthetic data generation for rare jaw lesions, CBCT artifact reduction, modality translation, image enhancement, and prognostic simulation. This chapter further examines evaluation strategies specific to generative models, emphasizing the limitations of pixel-level metrics and the importance of structural, distributional, and task-oriented validation. Safety considerations, hallucination risk, uncertainty modeling, and regulatory challenges are also addressed to ensure responsible clinical integration. By moving beyond discriminative analysis toward image generation and transformation, generative AI has the potential to augment diagnostic reliability, improve image quality, and support personalized decision-making in oral and maxillofacial radiology.