by Hayley F.
The reach of artificial intelligence is expanding beyond previously unthinkable domains and into some of the most inherently human areas of existence, such as medicine, interpersonal connection and even the domain of empathy. The narrative medicine movement, rooted in the concept of listening to patient stories and understanding disease as a lived experience, finds itself in the somewhat paradoxical position of navigating the benefits and unease generated by the growing capabilities of generative AI in health care. From summarizing patient charts, to dictating clinical notes to producing patient-simulated clinical conversations and even artificially emotional language, AI can do all of the above. These technologies offer benefits in terms of efficiency and access; however, there is the underlying question, what happens to patient narrative when stories are no longer processed purely by human beings and instead run through algorithmic systems?
The origins of narrative medicine stemmed directly from a response to the de-personalization of medicine, asserting that disease cannot be viewed solely through the prism of symptomology and biological deviation. Rather, it argued that illness must be examined in its contextualized totality; that it involves not only biology and symptoms, but also the fears, identities and histories that surround a patient. The patient story took center stage because stories are complex and full of information that a laboratory value cannot encapsulate. Feelings of pain, suffering, grief, fear and stigma exist in the human experience, though they often extend beyond clinically quantifiable health indicators.
Generative AI enters narrative medicine, potentially as a collaborator. With overworked doctors bogged down by documentation requirements and facing ever-greater time constraints in their interactions with patients, physicians have been using an increasing number of “ambient AI scribes,” technologies that listen to physician-patient interactions and automatically record the encounter and create clinical notes. Supporters of such technology contend it reduces physician administrative workload and exhaustion, allowing doctors to be present and attentive with patients, thereby expanding the realm of the empathic encounter, not diminishing it.
But an equally significant concern is the unintended flattening of human stories by technology. Patient narratives are rarely neat or seamless; they are characterized by silence, ambiguity, contradictions and breaks in thought and emotion. An algorithm optimized for clarity and speed might easily turn lived experience into clinically appropriate reports but would remove the nuances that make the stories meaningful to the human patient; an AI can turn the narratives into bits of data rather than human encounters.
The ethical implications extend beyond documentation. Recent case discussions in medical humanities have explored the use of AI-generated emotional support in healthcare settings, including systems designed to provide behavioral encouragement and emotionally responsive dialogue for patients managing chronic illness. While some patients report feeling heard and supported, these developments raise difficult questions about authenticity in healthcare relationships. A machine can imitate empathy convincingly, but it does not experience vulnerability, mortality, fear, or suffering itself.
Even more troubling are the recent case reports in psychiatry about “AI-associated psychosis.” In these reports, individuals who already had vulnerabilities found their thought processes altered and formed an unhealthy attachment to conversational AIs through increased interactions with them. These reports are a clear demonstration of the psychological power people are able to, and willing to give conversational AIs without even realizing it. The boundary between aid and artificial emotional connection is beginning to fade in the minds of patients who view conversational AIs as an emotional connection.
At the same time, rejecting AI entirely would oversimplify the issue. Technology has always mediated the human experience of medicine, through diagnostic and imaging tools to even earlier algorithmic models. The question is not whether or not AI has a place in medicine, but how its presence will be managed without destroying the humanistic aspect of narrative medicine that so many clinicians seek to protect. It should function as a tool rather than a replacement; the problems arise when efficiency and accuracy become more important than presence and when stories are processed for their system-appropriateness rather than their value to human experience.
Ultimately, the intersection of AI and narrative medicine reflects a broader tension within modern healthcare: the conflict between technological advancement and human connection. If medicine continues to digitize, then cultivating empathy, listening, and patient-centered care becomes even more critical, and the future of narrative medicine may very well be not an outright rejection of the technology itself, but its disciplined and thoughtful implementation in relation to human stories.
Art by Stockcake, “Medical Innovation“