This chapter provides an in-depth examination of one kind of generative AI model: diffusion models, which produce realistic data by transforming noise into structured outputs. It outlines the operation of diffusion models, including forward and reverse processes, and highlights their advantages over other generative models, such as GANs. It covers the mechanics of different types of diffusion models, including discrete, continuous, and hybrid models, as well as specialized forms, such as score-based models and denoising diffusion probabilistic models (DDPMs). Further, the chapter reviews various types of diffusion models and their use in fields such as digital art, animation, healthcare, and scientific research. A key focus is placed on the use of diffusion models in histopathology, where they enhance medical imaging by reducing noise and improving diagnostic accuracy. By generating clear, high-quality images, diffusion models support more precise diagnoses and contribute to advances in disease detection and treatment. The future of diffusion models is also discussed, including potential improvements and expanded applications across different industries.