Perspective
AI in Medicine
The Lost Aura of the Physician in the Age of Artificial Intelligence
John D. Lantos
JAMA Published Online: March 2, 2026
doi: 10.1001/jama.2026.0946
Artificial intelligence (AI) now performs many tasks that were once the exclusive province of physicians.1 It makes difficult diagnoses, provides psychological counseling, detects drug interactions, reads images, predicts outcomes, and reviews scientific articles (eAppendix in the Supplement).2 As these capacities expand, physicians’ roles will change. In many settings, physicians are increasingly positioned as supervisors of semiautonomous systems, retaining responsibility with diminished autonomy.3
Philosopher Walter Benjamin lived long before the advent of AI but he analyzed another transformative technology. Writing in the 1930s about photography and cinema, he asked whether the mechanical reproduction of visual images would render painting—and painters—obsolete. In Benjamin’s time, some critics denigrated the artistic value of photographs just as today many question whether AI can truly understand illness or empathize with discomfort.4
Benjamin recognized that technologic reproductions were different. He hypothesized that original works of art had a so-called aura. Benjamin never precisely defined aura. Instead, he recognized that there was something special about an original artwork compared with a reproduction, that it reflected its singular history and unique trajectory through time, space, and social meaning.5
Historically, physicians possessed something comparable. Their professional authority was grounded in the unique circumstances of their training, the practical wisdom that they had accrued, their face-to-face presence with patients, and their nuanced clinical judgment. Like an original painting, medical expertise appeared singular and inseparable from the clinician who exercised it rather than from the tools or institutions that supported the physician’s practice. But the reproductions are becoming increasingly convincing.
When writers doubt that a robotic physician could ever replace a human clinician, they emphasize humanistic qualities, such as empathy, accountability, and trust,6 and argue that machines can only simulate inadequate replicas of these essential qualities. The limits of telemedicine offer a cautionary paradigm. Telemedicine was once touted as a technology that would lower costs and expand access to expertise. Historian Jeremy Greene has shown that although many telemedicine initiatives succeeded on their own terms, they exposed a persistent limitation: in Greene’s words, “telepresence was not the same as presence.”7 Seen through Benjamin’s lens, face-to-face encounters between healer and patient appeared to possess something like an aura that made them uniquely human. Similar concerns may arise as AI comes into more widespread use.
The tension between the quest for technologic efficiency and insistence of the importance of human presence long predates AI. Those 2 goals have always been in tension. Every push in one direction leads to pushback in the other.8 Artificial intelligence is but the latest development in this long struggle over the soul of medicine.
Foucault and the Clinical Gaze
That long struggle began as early as the late 18th century. Until then, illness was understood largely through patients’ subjective experiences. The physician’s primary diagnostic tool was the medical history. Then, innovative French physicians started to reimagine disease as an objectively observable pathologic entity. The patients’ understanding of their ailment became extraneous.
Philosopher Michel Foucault analyzed this shift.9 He showed how, in the minds of these pioneers, individual physicians and patients were increasingly framed as “disturbances that can hardly be avoided” but that, ideally, should be neutralized in the pursuit of objective medical truth. Foucault called this abstraction the clinical gaze. The patient’s story and the physician’s interpretive presence became secondary to precise measurements and visualized pathology, quantification, and standardization. Physicians were encouraged to spend less time talking to patients and more time gathering the data that would lead to an evidence-based diagnosis and treatment plan.
Anesthesia, the Silent Body, and Hospital Medicine
The rise of modern surgery led to a further transformation in the patient-physician relationship. In the mid-19th century, anesthesia made surgery tolerable and technically possible by eliminating pain and movement, whereas antisepsis made it safer. Anesthesia also rendered patients insensible, motionless, and silent, allowing physicians to act on bodies without engaging subjectivity and permitting surgeons to temporarily set aside the patient as a person and repair the body as an object. Physicians had to adapt psychologically by compartmentalizing compassion and empathy, reframing emotional distance as professional competence rather than moral loss.10
By the early 20th century, this posture extended beyond the operating room. Historians Charles Rosenberg11 and Joel Howell12 show how the modern hospital organized clinical work around technologies and administrative practices that render patients the passive recipients of therapeutic interventions. The shift from home to hospital care meant that efficient medical treatment often came at the expense of deep psychosocial knowledge and trusting relationships. Francis Peabody worried that hospitals “deteriorate into dehumanized machines” in which “one of the first things that commonly happens to him is that the patient loses his personal identity.”13 Improved safety and effectiveness came at the cost of disrupted continuity and depersonalization.
Evidence-Based Medicine and the Algorithmic Clinician
The movement toward evidence-based medicine in the late 20th century accelerated these trends. Evidence-based medicine privileged recommendations grounded in reproducible evidence about population-level outcomes over individualized clinical judgment. This shift reflected legitimate ethical concerns. Unwarranted practice variation often led to unnecessary, potentially dangerous procedures.14
Evidence-based medicine redefined clinical expertise as the ability to apply generalizable evidence to individual cases.15 This reconceptualization proved especially compatible with emerging digital infrastructures. Clinical guidelines, decision rules, and quality metrics—central to evidence-based medicine—were readily translated into algorithms and then embedded into electronic health records.16
Over time, electronic health records, initially designed for documentation and billing, reshaped clinical thinking and clinical practice. Physicians began spending more time communicating with their computer screens than with their patients. Clinical findings and insights were carefully curated and coded in ways aligned with reimbursement structures. Relative value units, developed to standardize reimbursement, conceptualized clinical care as a series of discrete, measurable tasks rather than as an integrated relational practice. Clinicians learned to think in templates, checkboxes, and fields, not by preference but by necessity.17 Longitudinal outpatient care preserved some continuity but also became increasingly governed by documentation requirements, quality metrics, and computer-initiated prompts.18
By the time AI systems began generating notes and recommendations, physicians had been trained to think in ways compatible with machine logic. As sociologists Stefan Timmermans and Marc Berg observed, these systems transformed professional work by redefining what counts as knowledge, competence, and good practice.19 Artificial intelligence did not initiate this transformation; it arrived to perfect it.
By the early 21st century, clinical reasoning was encoded in protocols, decision rules, and electronic prompts. Professional judgment shifted from originating decisions to managing exceptions, adjudicating among machine-generated outputs while operating under institutional pressures not to deviate. Professional autonomy steadily eroded. Individualized clinical judgment and discretion became a source of risk.
How AI Is Different
Although AI is the logical outgrowth of these trends, it differs from prior technologic interventions in 2 key respects. First, AI is interactive. It can perform many nontechnical tasks that were once thought beyond the reach of technology and machine learning. It now obtains informed consent, counsels patients about end-of-life decisions, and provides psychotherapy. It does so with language that expresses attentiveness and concern.20 Perceived compassion and patient explanation no longer depend solely on the capacities of individual clinicians.21These capacities of AI might allow it to fulfill the dreams of both biotechnical medicine and humanistic medicine. Second, AI is available to all, on their phones, to be used without the need for an expensive superstructure like a hospital. Patients and physicians can innovate below the radar of traditional regulatory mechanisms.
When a profession’s core competencies become reproducible, the central question is not whether it will disappear but how its social role will be redefined. Physicians’ aura—once grounded in the irreducible presence of one human before another—will diminish as caring becomes technologically reproducible and warmth can be simulated on demand. The big question for physicians today is how they will reimagine their roles when many of the skills for which physicians were uniquely valued are now reproducible at scale, available to all, and constantly improving.
In Benjamin’s account, photography and film produced the shock of the new by stripping art of its uniqueness, ritual authority, and distance. Yet the loss of aura did not mark the end of art or artists. Freed from the obligation to reproduce reality, painters and other artists reinvented their social role, exploring new forms of perception, expression, and critique. Impressionism, abstraction, and conceptual art emerged. The shock of the new forced a reckoning with the meaning of art.
For medicine, such a reckoning will require bold thinking about which skills are uniquely human, which can be replaced by sophisticated technology, and what new ideas of clinical excellence might emerge. Widely agreed-on standards for how physicians are educated, how their skills are evaluated, and what counts as excellence will likely be challenged. The disruption of art led to deep controversies about what art is, what it should do, and how to judge quality. Artificial intelligence’s disruption of medicine will likely lead to similar discussions about physicians, medicine, and the work of healing.
In Awakenings,22 Oliver Sacks recognized that “all of us entertain the idea of another sort of medicine…which will restore us to our lost health and wholeness.” Perhaps that is now within reach. Whether physicians can dare to reinvent their social role and professional identity is an open and urgent question.
