Ars Longa, Data Brevis: the science of disease and the art of care

The integration of Artificial Intelligence (AI) into modern medicine has revolutionised diagnostic accuracy, yet it generates a critical ethical dilemma: as healthcare becomes more data-driven, it risks eroding the high-touch essence of care. As algorithms increasingly shape clinical decision-making, patients risk being reduced to data points rather than persons with unique life stories. This paper examines the tension between AI’s calculative logic and the narrative nature of illness, introducing the Processing–Witnessing Model. This framework distinguishes between algorithmic processing (speed, optimisation) and human witnessing (presence, interpretation, and narrative understanding). While AI excels at managing disease as biological dysfunction, it cannot address illness as the lived experience of suffering. This paper argues that the opacity of black-box systems creates a contextual void, enabling a form of epistemic violence that renders the patient’s story invisible. Furthermore, the normalization of the screen gaze threatens the therapeutic alliance. Ironically, studies rating AI chatbots as ‘more empathetic’ are interpreted here not as evidence of machine moral agency, but as a symptom of systemic burnout. This makes narrative ethics a priority for medical education. The physician’s role should evolve from data supervisor to cultural mediator, ensuring that while machines process the biological hardware, clinicians provide the interpretive software of meaning.