When the lab launches later this month, 80 Dell Med clinicians — a mix of faculty, residents, and students — will meet with patients in simulated encounters.
Published October 20, 2026
The IC² Institute Institute and Dell Medical School are launching an innovative research initiative examining the impact of AI on doctors’ clinical reasoning skills. Through a series of simulated clinician/patient encounters designed to yield empirical data, the AI Learning Lab will help researchers understand how doctors are using artificial intelligence to support their own knowledge, judgement and reasoning.
Study results could help Dell Med and other institutions refine their approach to how and when they incorporate AI into their curriculum.
The Learning Lab is timely: Dell Med is now the anchor of UT’s emerging, AI-native academic health system. And a 2026 American Medical Association survey found that 81% of doctors are using AI in their practice — more than double the 38% reported in 2023. With the steep rise in AI use among practicing doctors, medical schools are grappling with the question of how and when to introduce AI to their students and trainees.
James Korndorffer, M.D., Vice Dean of Education for Dell Med, explained: “We know that AI is going to be, and should be, used. This research coming out of the AI Learning Lab may give us insight into where these tools might create gaps, and when to introduce AI tools for our learners in a way that supports the best patient care possible.”
The Learning Lab in Action
When the lab launches later this month, 80 Dell Med clinicians — a mix of faculty, residents, and students — will meet with patients in simulated encounters. The patients (paid actors) will present with a medical issue; the clinicians will meet with the patient, ask questions, synthesize information, and arrive at a diagnosis.
Half of the clinicians will have Claude at their fingertips to assist with diagnosing the patient; the other half will be reliant on UpToDate, a web-based medical encyclopedia — meaning no AI. Researchers will observe and compare the clinical reasoning processes of both groups and also analyze the use of AI across experience levels.
A Different Look at AI in Medicine
The Learning Lab goes beyond recent studies of clinically applied AI in two key aspects. First, most other studies are simply measuring the performance of AI — essentially asking, “Can AI get to the correct answer? And who can get there faster — AI or human clinician?”
The AI Learning Lab is aimed at a different, more detailed, set of questions. Researcher Sookja Kang lists a few of them: “How quickly do the clinician participants click on their AI tool? Are they passively accepting AI output or actively interrogating it? How much reasoning work gets handed off to the AI tool, especially as the diagnosis evolves?”
With these sorts of questions, researchers hope to capture the details around how and when AI tools are being used and how, precisely, AI tools may be reshaping doctors’ clinical workflows. To get their answers, the team will be deploying an arsenal of measurements and rubrics as well as observational studies and an analysis of post-encounter interviews.
The second way in which the Learning Lab differs from other studies is the use of a live patient encounter as opposed to a static text vignette.
Researcher TuQuyen Dao explained why the live encounter is important: “A large part of solving a real clinical case is interacting with the patient to gather information, build and revise the problem representation. Our study more closely resembles a real-world clinical scenario where information is revealed dynamically through patient interaction.”
Examining AI’s Relationship to Clinical Reasoning Skills
The Learning Lab extends previous IC² Institute explorations of the impact of AI on doctors’ clinical reasoning. A recent qualitative study suggests that rather than diminishing cognitive engagement (a common concern about AI across many fields), AI can actually push doctors to develop a new set of skills. Similarly, the IC² Institute research suggests that AI-generated outputs may expand doctors’ thinking about possible diagnoses and treatment plans. The learning lab is a vehicle for understanding those dynamics using empirical data.
For educators at Dell Med, the lab represents an opportunity to better understand how both faculty and learners are interacting with AI. That information could help ensure that AI is introduced responsibly and effectively during training, with the end goal of preparing doctors to deliver the best patient care possible.
LEARN MORE
AI Learning Lab: Understanding the Impact of Large Language Models on Clinical Reasoning