As one of the largest breakthroughs in AI, ChatGPT has attracted tremendous attention and been applied to various domains. Existing applications include direct applications of ChatGPT, fine-tuning the GPT model by domain-specific data, and developing a ChatGPT-like framework for specific tasks. In this talk, I will demonstrate how ChatGPT-like frameworks can be developed to solve open problems in biomedicine, through the introduction of some of our recent works. In particular, I will introduce our work on predicting the evolution of SARS-COV-2 virus, on designing antibodies with simultaneous optimization of multiple metrics, on SkinGPT4 which is the first multimodal dermatological diagnosis framework, and on PepGPT, a ChatGPT-like framework that can design peptide drugs through interaction with users.