Why Artificial Intelligence in Medicine Cannot Replace Clinical Intuition
By Garrett Terracciano, MD | May 2, 2026
On a cold January night, I worked a 34-hour shift as a senior pediatric resident in a large children’s hospital in the Northeast. Four new admissions came one after the other. One patient was a teenage boy who had new-onset type 1 diabetes and showed signs of diabetic ketoacidosis (DKA). His lab numbers improved. His anion gap had closed. On paper, he was getting better and ready to move to a lower level of care.
I went to the emergency department to see him for myself. I wanted to support my busy intern. The boy was awake and his chart said he was “alert and oriented times three.” Still, something was off. His eyes were glazed. His speech was slurred. He sat too still. My clinical intuition, built over many years, told me that something was very wrong.
I faced pushback from the pediatric intensive care fellow and the emergency department attending. They said he was just tired from waiting. But I trusted my senses and cited known medical standards. I warned about risks like cerebral edema from correcting DKA too quickly.
The Limits of Medical Documentation
Every time a doctor leaves a patient’s room, we try to capture the moment in a chart. This chart condenses a rich human encounter into labs, numbers, and written notes. Much of the sensory detail and subtle signs do not make it into the record.
The philosopher Michael Polanyi spoke of “tacit knowledge.” We know more than we can say. In medicine, experienced doctors build a detailed mental model of disease. They use patterns and all the subtle cues from body language, tone, and movement. This depth does not appear in lab reports or notes.
AI systems, including large language models, learn only from written records. They see text and the statistical ties between words. They miss the patient’s presence. As a result, AI creates a “language model” without the full “world model” that doctors build from watching and listening.
Why AI Cannot Fully Replace Clinical Intuition
Clinical AI tools learn from the chart, a short version of the patient encounter. This means they may ignore nuances and non-verbal details that are key to care.
We already see these limits. For example, a language model used in a health tool was studied by Nature Medicine. This tool under-triaged more than half of true emergencies. It wrongly suggested that patients with DKA only needed routine follow-up. The algorithm found danger in its own reasoning. Yet it reassured the patient based only on the chart data.
This incident shows the truth. AI can often identify diagnoses by following patterns. But it lacks the full understanding of a doctor. Its strength in pattern matching does not replace the depth of true clinical insight.
The Risk of Overreliance on AI in Clinical Practice
AI brings great power to healthcare. It finds patterns in large groups and can help improve outcomes. But the hidden risk is its quiet success. When doctors trust charts and algorithms more than their own senses, critical cues may be lost.
Workflows may change. More emphasis can be placed on structured data for AI use. This may force doctors to focus on rigid, written entries. In this process, the unstructured details that doctors notice over years of practice may disappear. The art of listening to a patient’s story and subtle signals may be slowly erased.
Conclusion: Preserving Clinical Intuition in the Age of AI
Before morning signout, I walked to the ICU. There, I saw the boy from earlier. A makeshift sterile field surrounded him. He was intubated and sedated to protect his swollen brain. His numbers had seemed to improve, but his body told a very different story.
This experience offers a clear warning to the medical community and AI developers. AI cannot replace clinical intuition. It misses the lived experience and raw perception of a doctor. The key question remains: will our health system keep the skill of “hearing” the quiet signals that live beyond a chart?
AI is a powerful tool. But it should support, not replace, the human ability to listen, sense, and act on the unspoken. This balance lets medicine honor both the art and the science of healing.
Garrett Terracciano, MD, is an anesthesiologist and medical writer. This article originally appeared on KevinMD.
Related Reading:
- Artificial intelligence in clinical care: Shaping the HHS policy landscape
- Evidence-based medicine vs. clinical judgment: a medical student’s perspective
- From penicillin to digital health: the impact of social media on medicine
For more insightful physician commentary, visit KevinMD.com.
Try this workflow today, Writer Link AI and Write Easy provide smart outputs with a natural voice. Get started with a free plan at