In The Birth of the Clinic, Michel Foucault introduced the concept of the “medical gaze” to describe how physicians translate a patient’s lived experience into biomedical terms, emphasizing information that fits a clinical framework while setting aside what does not. Every gaze involves selecting certain details from the vast stream of sensory information. In medicine, that selection often privileges disease, symptoms, and measurable findings while overlooking the personal, social, and emotional dimensions of illness, partly because the biomedical perspective is more familiar and manageable.

Thomas Eakins’s surgical canvases, most famously “The Gross Clinic,” capture the essence of the medical gaze. The physician observes, the student watches, and the patient’s exposed body becomes the site of knowledge production. Yet the paintings also reveal that observation in medicine is never neutral. Who is seen, who is permitted to look, what is exposed, and how the image is used are all shaped by power, pedagogy, and culture.

Beyond the medical gaze lies an ethical gaze: a way of seeing that holds observation accountable. It questions not only whether an image is clinically useful but also whether its creation, display, and reuse respect consent, privacy, representation, and personhood.

Observation, history, and hidden bodies

Modern anatomical education was built, in part, on the bodies of people who had little or no control over what happened to them after death. European and British dissection relied heavily on executed prisoners and the unclaimed dead from workhouses, hospitals, prisons, and asylums. Their bodies entered classrooms and atlases, while their names and histories often disappeared.

Pernkopf’s atlas is among the clearest examples. Its technical brilliance cannot be separated from the bodies of people executed under the Nazi regime, including political prisoners and resistance members. William Hunter’s celebrated plates of the gravid uterus raise a related, though historically more contested, question: How were rare and intimate bodies obtained and transformed into teaching objects in an era when consent was not the governing norm?

This history makes provenance more than an archival detail. Whenever medicine uses powerful images, educators should explain where they came from, whose bodies made them possible, and which inequalities shaped their visibility or anonymity.

Representation, power, and who is seen

Medical images do more than depict disease. They also teach learners what a “normal” body looks like and whose body serves as the default. Many atlases and teaching sets have centered on light-skinned, idealized figures, while socially marginalized bodies appear primarily as anonymous examples of pathology.

In dermatology, this imbalance has direct clinical consequences. Diseases may present differently across skin tones, yet darker skin remains underrepresented in many educational resources. A learner trained primarily on images of light skin may be less prepared to recognize disease in patients with skin of color.

Ethical observation therefore requires active inclusion. Teaching collections should represent a broad range of skin tones, body sizes, ages, sexes, gender identities, and physical differences, not as a gesture of diversity, but as a requirement for diagnostic competence and equitable care.

Consent, privacy, and the intimacy of images

Historical images often provide no evidence that the person depicted consented to being photographed, dissected, or displayed. Contemporary technology does not eliminate that problem; in some ways, it exacerbates it. A clinical photograph can be captured in seconds, copied exactly, stored indefinitely, and circulated far beyond its original purpose.

Consent should therefore be specific rather than assumed. Patients should understand why an image is being created, who may view it, how long it will be retained, whether it may be used for teaching, publication, research, or legal proceedings, and whether future uses can be meaningfully limited. Deidentification is important, but it does not always erase an image’s intimacy, or even its recognizability.

Autopsy and forensic images require especially strict stewardship. They may be medically or legally necessary, but necessity does not render them ethically weightless. The dignity of the person depicted should endure after death.

Objectification versus personhood

An image does not merely record a patient; it frames the patient. Cropping, labeling, posture, lighting, and context can reduce a person to a specimen. Once the face and story disappear, it becomes easier to forget that the teaching image depicts someone who experienced fear, pain, shame, hope, and trust.

Weight-related images illustrate the problem. They can reinforce stigma by framing body size as a spectacle or a personal failure. Used thoughtfully, however, art and the medical humanities can help clinicians question culturally contingent assumptions about beauty, health, discipline, and blame.

Museum-based programs offer a practical model. By asking trainees to separate description from interpretation, these programs teach them to distinguish what they observe from what they assume. Such training can improve descriptive observation and may strengthen perspective-taking. More importantly, it reminds clinicians that every act of looking is interpretive.

Digital and AI images: authenticity, bias, and responsibility

AI-generated medical images may offer scalable, customizable teaching materials, especially in visually driven specialties. They can illustrate rare conditions, provide varied examples, and reduce some privacy concerns associated with real patient images.

But synthetic images pose new ethical risks. They may contain subtle clinical inaccuracies, reproduce biases in datasets, present unrealistic combinations of findings, or circulate without clear labels. As photorealism improves, learners may not know whether an image is authentic, altered, or entirely synthetic.

The response should be governance, not prohibition. Synthetic images should be clearly labeled, clinically validated, traceable to their source, and curated by qualified educators. AI literacy must include not only how to generate images but also how to question them. Synthetic material should augment, not replace, contact with real patients and the independent visual judgment that clinical practice requires.

Teaching an ethical gaze

Teaching about bodies entails several obligations. Clinicians and educators should disclose provenance, acknowledge historical injustice, obtain meaningful consent, protect privacy, diversify representation, and teach learners to distinguish observation from assumption. They should also establish standards for creating, validating, labeling, storing, and reusing digital and AI-generated images. The goal is not to scrutinize medicine less closely. It is to scrutinize medicine more responsibly.

Eakins’s surgical theaters anticipated a continuing truth: The patient’s body makes clinical knowledge possible, yet usefulness does not erase personhood. Every act of looking carries ethical weight. Physicians must learn not only to see accurately but also to see responsibly.

Arthur Lazarus is a physician-author whose work spans narrative medicine, physician leadership, artificial intelligence, health care ethics, medical culture, and fiction. He has published more than 500 articles and essays across scientific journals, professional publications, and online platforms.

He is the author of numerous books on narrative medicine, AI in medicine, career development, and the changing moral landscape of health care, as well as fictional series including Rounds Never End, Sick and Systemic, and Real Medicine, Unreal Stories. His writing explores the forces reshaping modern medicine while preserving a central commitment to story, meaning, judgment, and the human relationship at the heart of care.

He shares updates on LinkedIn.