Abstract The integration of artificial intelligence into clinical practice demands fundamental reconsideration of how we conceptualize diagnosis. Clinical and popular discourse often frames diagnosis as a “virtuoso model” emphasizing individual physician expertise, which inadequately addresses the complexities of human-AI collaboration and overlooks work on its distributed and social dimensions. We propose a pluralistic framework moving beyond this paradigm toward a model better suited for AI integration. Drawing from medical and cultural history, sociological analysis, and cognitive science, we identify four key shifts necessary to reconceptualize diagnosis for the AI context: recognizing diagnosis as distributed cognition rather than simply individual virtuosity; embracing metacognitive awareness for human-AI collaboration; acknowledging patients as diagnostic co-creators rather than passive subjects; and understanding diagnosis as contextual and culturally situated. Successful AI integration requires moving beyond viewing AI as threat or efficiency tool toward reconceptualizing diagnosis as a collaborative, technologically-mediated activity where clinicians function as interpreters, patients as co-creators, and AI as specialized partners with transparent limitations. This framework demands new approaches to medical education and human-computer interaction research to realize AI’s potential while preserving essential human judgment.
Krishnan et al. (Thu,) studied this question.