[Applications and challenges of large language models in critical care medicine].

The rapid development of big data methods and technologies has provided more and more new ideas and methods for clinical diagnosis and treatment. The emergence of large language models (LLM) has made it possible for human-computer interactive dialogues and applications in complex medical scenarios. Critical care medicine is a process of continuous dynamic targeted treatment. The huge data generated in this process needs to be integrated and optimized through models for clinical application, interaction in teaching simulation, and assistance in scientific research. Using the LLM represented by generative pre-trained transformer ChatGPT can initially realize the application in the diagnosis of severe diseases, the prediction of death risk and the management of medical records. At the same time, the time and space limitations, illusions and ethical and moral issues of ChatGPT emerged as the times require. In the future, it is undeniable that it may play a huge role in the diagnosis and treatment of critical care medicine, but the current application should be combined with more clinical knowledge reserves of critical care medicine to carefully judge its conclusions.