BackgroundVirtual reality simulations are shown to be an effective approach for interprofessional nurse-physician communication training. However, its scalability is constrained by unequal medical-nursing cohort size, rendering a great challenge for all nursing students to form an interprofessional team with medical students. With the evolution of artificial intelligence (AI), an AI medical team player can be integrated into virtual reality simulations for more nursing students to engage in interprofessional team training. ObjectivesTo describe the development of a novel AI-enabled virtual reality simulation (AI-enabled VRS) and to evaluate nursing students' competencies and experiences in communicating with an AI medical doctor. MethodsA mixed-methods design using a one-group pretest-posttest design and focus group discussions were employed in the evaluation phase. Nursing students from a university were recruited to undertake the 2-hour AI-enabled VRS. Pre-test and post-tests were administered to evaluate the participants' communication knowledge and self-efficacy. Survey questionnaires were administered to examine their experiences with the virtual reality environment and the AI doctor. Five focus group discussions were conducted to gain deeper insight into their learning experiences. ResultsThe participants demonstrated significant improvements in communication knowledge and interprofessional communication self-efficacy after the learning. They reported positively on the acceptability, feasibility and usability of the AI-enabled VRS. The subscale of “human-like” feature of the AI medical doctor was rated the lowest. Three themes surrounding participants' experiences of the virtual learning emerged: “relate to the real world”, “artificial intelligence versus human intelligence” and “complement with face-to-face learning”. ConclusionsThis study demonstrates initial evidence on the potential of AI-enabled VRS in fostering nursing students' learning on interprofessional communication skills. The findings have also provided insights on how to improve the AI-enabled VRS, in particular, the expressiveness of the AI pedagogical agent and facilitating more dialogue trainings with learner-agent conversations.