Rapid Review of Generative AI in Smart Medical Applications

Artificial intelligence, as a foundational technology in digital construction and smart healthcare, is significantly advancing the development of smart healthcare systems. The emergence of large AI models such as DALL-E, GPT-4, and LLaMA has brought unprecedented technological breakthroughs[5]. These large models, also known as pre-trained or foundation models, can centralize multimodal data, pre-train on extensive datasets with ultralarge-scale parameters, and fine-tune for various domain-specific tasks [12]. In healthcare, where medical data is inherently multimodal, large models are poised to accelerate smart medicine, the medical metaverse, and medical research. Applications range from electronic medical record comprehension, medical Q&A, education and training, image generation, disease diagnosis, drug development, to virtual interactions with hospitals and digital human avatars [43], covering all stages of medical care.