Why Handwritten Medical Notes May Still Matter in the Age of AI Scribes

Ambient artificial intelligence scribes are among the most rapidly adopted technologies in modern medicine, promising to relieve clinicians of the documentation burden that has become one of the leading drivers of burnout. But a new perspective published in the Journal of General Internal Medicine argues that the rush to automate note-writing may come at a hidden cost: the erosion of a practice that has quietly shaped the moral and cognitive development of physicians for millennia. The paper, authored by Joshua Briscoe and colleagues at Duke University School of Medicine and the Kennedy Institute of Ethics at Georgetown University, contends that note-writing is not merely a summative task of recording information but a formative one that cultivates the habits of attention, reasoning, and judgment that define good clinical practice.

The authors observe that current discourse surrounding AI scribes has focused overwhelmingly on questions of privacy, accuracy, and efficiency. These are important concerns, but they frame note-writing as primarily a communicative function—something that can be offloaded to a machine without consequence. Briscoe and his co-authors argue that this framing misses something fundamental. Writing, they suggest, is not just how clinicians convey what happened in a clinical encounter; it is how they learn to think clinically and morally. The act of composing a note requires a physician to attend carefully to what matters in a patient’s story, to organize and interrogate their own reasoning, and to justify their decisions in a way that can be scrutinized by others. These are not incidental byproducts of documentation. They are, the authors argue, central to the formation of a physician capable of caring well for patients.

The historical evidence for this claim is striking. Extant documents from the pre-Hippocratic era in the ancient Near East already contain patient profiles, therapies, and clinical outcomes, allowing physicians to learn from prior cases and teach apprentices. Hippocratic medical note-writing emphasized the importance of delineating clinical reasoning and justifying decision-making as a mode of self-pedagogy. From its earliest days, clinical documentation has served both summative and formative purposes. Summative writing justifies an assessment and plan; formative writing helps clinicians think by forcing them to interrogate and discover new ideas. Even as daily clinical documentation became standard practice in the eighteenth and nineteenth centuries, note-writing maintained its role not merely in communicating information but in organizing the thinking of the clinician and educating trainees.

This philosophy was perhaps most clearly articulated by Larry Weed, the pioneer of the SOAP note, who argued that the purpose of structured documentation was not strictly for communication but for the synthesis and crystallization of one’s own thoughts. It is no accident, Weed insisted, that the Assessment and Plan sections organically follow the Subjective and Objective sections: the structure itself mirrors the cognitive work of clinical reasoning. In this way, note-writing has historically played a central role in the development of clinical judgment, not just the conveyance of facts.

The advent of the electronic medical record promised improvements in data accuracy, continuity of care, and efficiency. Yet the perceived efficiency gains have increasingly drawn clinicians away from the bedside. Time and motion studies have shown that non-surgical physicians now spend a substantial portion of their day engaged with the EMR, either reviewing data or writing notes. Decentralizing the medical record through digitalization has allowed other stakeholders—from administrators to payors—to increase their influence over how it is used. Physicians now spend much of their time writing to justify billing, defend against legal liability, and account for metrics that may have little direct relevance to the specific patient in front of them. Pressure to accommodate these many stakeholders has resulted in notes bloated with extraneous, clinically irrelevant information and out-of-date copy-and-pasted text, rendering note-writing less a clinical or pedagogical activity than a bureaucratic box to check. Documentation has consequently earned its place as one of the main drivers of burnout in contemporary medicine.

Against this backdrop, the appeal of AI scribes is entirely understandable. Preliminary investigations suggest significant benefits in efficiency and accuracy, with the added promise of burnout mitigation. Human scribes have increased productivity and improved clinician satisfaction but at the expense of adding another staff member to the team. AI scribes seem to meet the eager needs of clinicians without incurring those costs. Yet the authors of the new perspective warn that in moments when many agree on the promise of a new technology, we are less likely to notice the potential benefits we leave behind. None of the existing critiques of AI-enhanced note-writing, they note, has grappled with a more fundamental question: what is the relationship between writing and medicine? And its corollary—what happens to the physician when they cease to write?

To answer this question, the authors draw on the philosopher Shannon Vallor, who describes seven mutually reinforcing aspects of moral formation: habits of mind, relational understanding, reflective self-understanding, habituation, modeling, perceptual attention to moral salience, and appropriate extension of moral concern. Because these facets reinforce one another, reflection and change may begin from any one of them. A physician might critically reflect on their documentation habits and whether those serve or detract from their practice of good medicine, or they might consider what they owe their patients in terms of their documentation. Writing, in this framework, is not just a tool but a practice through which clinicians cultivate the attentional and moral virtues that clinical care demands.

The authors also draw on Andrew Zelny’s exploration of four technological relations: technology can displace, diminish, augment, or define contexts for human agency and the exercise of wisdom. Early uses of AI scribes may diminish the physician’s role as writer in the clinical encounter, but they could soon displace the practice of writing and critical engagement with the medical record almost entirely. AI scribes could evolve to redefine the clinical encounter itself, either by further shaping clinician behavior to suit the metrics of systems that monitor and evaluate documentation, or by making the encounter more asynchronous—for example, an AI scribe tracking health data throughout a patient’s day to be sent to a clinician for review. Whether through total displacement or subtle diminishment, AI scribes may invite clinicians to offload habits of mind to automation, encouraging passivity in clinical reasoning and weakening important processes of moral and cognitive habituation. AI scribes may also sever physicians from one another as AI performs both the writing and the reading within the chart, thereby impairing relational understanding within a community of practice. As the formative act of writing is displaced or diminished, clinicians may have less opportunity for reflective self-understanding. Automation may thus diminish one venue in which clinicians can intentionally cultivate their own moral development and model such development to trainees. These dynamics may also increase the likelihood that clinicians will cede more of their judgment to automation—a particularly grave risk for trainee physicians who have not yet developed the cognitive habits that writing helps cultivate.

Yet the authors are careful not to dismiss the potential benefits of thoughtful AI use. Losing the tool of writing may create new opportunities for reflection, whether through additional time or through other forms of writing, such as a parallel chart where a clinician reflects on their experience of care. When AI scribes write for physicians, it may be possible for physicians to attend to salient factors of the clinical encounter without the distractions of tending simultaneously to the note itself, thus improving perceptual attention to moral salience. An AI scribe could also help clinicians notice aspects of a patient’s life that are easily overlooked in harried encounters, like the patient’s close relationships, living arrangements, and baseline functional status. But all of these potential benefits presume a mind already attuned to those important factors—one capable of resisting other distractions within the clinical encounter. They presume the attentional virtues needed to resist the dynamics of the attention economy, virtues honed prior to the deployment of AI technology. Moreover, they presume that time freed up by the use of AI scribes would in fact be allotted toward more careful attention to patients and their stories, rather than used institutionally to justify seeing more patients in less time. Because the drive for efficiency may override the best intentions of clinicians at the bedside, health systems must carefully consider the values and incentives embedded in the use of AI scribes.

The long-term formative effects of such technologies are difficult to anticipate. By the time adverse consequences become fully apparent, AI scribes may already be too entrenched within clinical practice—a challenge known as the Collingridge dilemma. The authors submit that attending to clinicians’ moral formation provides a response to this dilemma, not by anticipating every risk in advance, but by cultivating clinicians capable of exercising practical wisdom amid changing technological conditions. If AI scribes prove effective, pressures may increase to extend these systems beyond documentation toward interpretation, prioritization, and recommendation. As clinical reasoning is increasingly scaffolded by such systems, the physician’s role may shift from authoring judgments to authorizing them—the clinician’s role as human in the loop may involve little more than quality control on and attestation to automated processes. What begins as a relief from the burdens of writing may culminate in the quiet reconfiguration of professional agency. The process of becoming a physician is necessarily arduous, and this difficulty is a feature of medical training, not a problem to be resolved as efficiently as possible. The question, the authors conclude, is not merely whether AI can assume the work of writing, but what happens to clinician judgment when physicians no longer write.

Subject of Research: Medical Ethics and Artificial Intelligence

Article Title: In Defense of Note Writing: Moral Formation and the Use of AI Scribes

Article References: Briscoe, J., Elmore, M., Tanzillo, D., & Frush, B. (2026). In Defense of Note Writing: Moral Formation and the Use of AI Scribes. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10771-2

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10771-2

Keywords: Defense, Note, Writing, Moral, Formation, Scribes, scientific research