Professional Identity Formation (PIF) is considered a crucial process in medical education. It involves how medical students identify their role as physicians, discover their professional positioning, and gradually develop their professional identity through social interactions. This qualitative descriptive study adopted the phenomenological method; it proposed the DSCOR (Diverse thinking, Seeking advice, Construction, Organizing and sharing, and Reflection) model based on the ChatGPT as a life coach (ChatGPT-LC) selfregulated learning (SRL) approach to design a PIF course for six medical students. Data collection included digital storytelling created using AI-generated techniques, learning sheets, direct observations, reflective feedback forms, and semi-structured interviews. The data transcription and analysis were conducted using Colaizzi's method. The results revealed three benefits of Generative AI (i.e., ChatGPT), namely "increasing motivation for planning PIF," "strengthening the mastery of PIF," and "broadening the perspectives of PIF." Moreover, the ChatGPT-LC SRL approach had a positive influence on students, helping them understand the significance of PIF in their personal development at the early stage. The artificial intelligence-generated content provided positive guidance and supportive learning, offering specific suggestions and assistance. This brought about benefits for learning, and provided initial evidence for the application of ChatGPT-LC in medical education.