<p>While on the surface the rapid adoption of generative artificial intelligence (AI) in healthcare signals a new technological innovation that may sideline the social sciences, arts, and literature comprising the medical humanities, we argue that one way to understand the applications of AI based on large language models (LLM) in healthcare is as a deeply narrative project. LLM-based AI endorses narrative and humanistic value insofar as it is trained on vast amounts of narrative data as inputs and generates outputs in narrative formats. We contend that the medical humanities, rather than being replaced by AI in healthcare, are all the more essential to understand the practical and ethical opportunities and constraints for responsible use and integration of such tools. By analyzing two case studies of generative AI in healthcare reported in literature, we show that narrative medicine and the medical humanities provide crucial theoretical frameworks and disciplinary skills to assess, integrate, and critique the roles and impacts of such technologies in healthcare. Rather than offsetting narrative practice via AI, responsible use of LLM-based AI in healthcare requires attention to the functions of narrative in medicine and the importance of the medical humanities.</p>