In the Age of AI, Bedside Physical Exam Skills Matter More Than Ever

July 10, 2026 - By Lindsay Paulsen

The proliferation of artificial intelligence (AI) has led many to predict that the technology will pull physicians further away from the patient bedside, resulting in a decline of the physical exam and the associated skills.

At Stanford Medicine 25, we believe this moment is an opportunity to elevate the physical exam. In fact, we’d like to make the case that the opposite is true: Bedside physical exam skills matter more than ever.

The Timeless Ritual of the Physical Exam

As a leader in medical humanities and an infectious disease physician, Stanford Medicine 25’s founder, Abraham Verghese, MD, has been a longtime advocate for the physical exam.

Verghese explains that, at its core, the physical exam at the patient bedside is a timeless ritual that marks a transformation. “It's the sealing of the patient-physician relationship,” says Verghese, in an article by Georgia Garvey for the American Medical Association. “It is the transition from ignoring a symptom to doing something about it, hopefully a transition from illness to wellness.”

In addition to being an important opportunity for human connection, interventional physiatrist Francisco M. Torres, MD, reminds us that the physical exam has practical benefits, which he describes in an article for KevinMD.com. It’s cost-free, non-invasive, immediately available and “capable of detecting a wide range of abnormalities that imaging may miss or misinterpret,” he explains.

Torres continues: “Our senses, sight, touch and hearing, are the product of millions of years of evolution. They are finely attuned to detect subtle changes in temperature, texture, symmetry, and movement. No imaging technology can match the breadth of information a skilled clinician can gather within minutes at the bedside.”

The Evolution of the Physical Exam and the Pull of Technology

Verghese, tracing the history of medical diagnosis, explains that most medical care prior to 1819 was carried out by barber surgeons who focused on cupping, bleeding and purging, rather than medical diagnosis. Later, the adoption of percussion and the invention of the stethoscope led to physical exams with the goal of diagnosis. Verghese emphasizes that, during this period, “the skills at the bedside reached incredible heights.” Physicians became highly skilled at reading the body.

Then, as medical technology advanced, introducing more sophisticated imaging and the electronic health record (EHR), physicians were pulled away from the bedside. “It’s very clear that a great part of our attention, a great part of our knowledge about the patient, is no longer on the patient’s body. It is actually in the computer,” says Verghese, pointing out the rise in diagnostic errors that occur simply because patients aren’t thoroughly examined.

It can also be argued that the necessary and inevitable advancement of medical technology actually places greater importance on the need for physicians to gather correct “data inputs.”

In an article for KevinMD.com, cardiologist Jason Ryan, MD, writes: “If we begin using AI extensively to support clinical practice, the most important role of the physician may become something surprisingly old-fashioned: performing an accurate physical exam. Physicians who excel at this form of data gathering may find themselves more valuable than ever.”

Ryan continues: “As anyone who has experimented with large language models knows, the outputs of AI are only as good as the prompts given. Medicine will be no different. The physician who reports clear lungs when subtle crackles are present in the right lower lobe will inadvertently generate poor AI recommendations. The doctor who documents a benign abdomen in a patient with early appendicitis introduces error before the algorithm even begins to reason. When exam inputs are wrong, downstream decisions are corrupted.”

A Physician’s Call to Action

At Stanford Medicine, located in the heart of Silicon Valley, we see great opportunity in technological innovation and its thoughtful applications to medicine and patient care. This is a critical moment in time that presents an opportunity for physicians to consider how AI can help bring us closer to the bedside, rather than push us further.

Our own Stanford Medicine 25 team members Poonam Hosamani, MD, Jeffrey Chi, MD, and Stanford colleagues Kevin Keet, MD, and Jason Hom, MD, capture this idea well in a 2026 editorial for BMJ Digital Health & AI: “If AI helps learners to reclaim time spent in front of the computer and returns us all to the bedside, that would be a triumph. Conversely, if the iPatient becomes an even more abstract ‘AI-Patient,’ where trainees spend even more time interacting with a computer than the actual patient, the consequences could be detrimental to future generations of patients and physicians alike.”