5 October 2026 · Medows
AI Scribes Cost Pre-Meds a Window Into Medicine
U.Va. Health cut student scribe shifts as ambient AI expanded. The note got easier. The apprenticeship got thinner. What that means for training.
A student scribe used to stand in the corner of the emergency department, laptop open, typing while a physician talked to a patient. Nobody called it teaching. It was teaching anyway.
On October 1, the student paper at the University of Virginia reported that this job is shrinking. The tool that is shrinking it is an ambient AI scribe.
What happened at U.Va. Health
According to The Cavalier Daily, U.Va. Health uses DAX Copilot, Microsoft's ambient tool that drafts notes from clinician-patient conversations. It was implemented in 2025.
In April 2026, the undergraduate scribe program's annual budget was cut by one-third. A new emergency department schedule began on September 15, 2026, reducing or eliminating some scribe shifts.
The paper names the effect on individual students. One fourth-year went from 16 to 20 hours a week to 20 hours for the whole month of October. Another dropped from two or three eight-hour shifts a week to about one shift of 4 to 5 hours. A third now works peak hours only, for example 10am to 3pm instead of a full 7am to 3pm resident shift. The scribes are paid $13 to $13.50 an hour.
One of them told the paper: "I'm getting less clinical experience, less income and less opportunities to learn."
The physician quoted in the piece, Dr. Sarah Wendel, said she did not make the budget changes, and that decisions reflected broader system-wide cost management. Dr. Richard Andrew Taylor raised a different worry: "deskilling and learning effective clinical reasoning."
The part that is easy to miss
The story is about a side effect, not about note quality.
Scribing was never designed as a pipeline. It became one. Pre-med students watched real histories being taken, saw how a senior phrases a hard question, heard how a differential gets narrowed in real time. They did it for $13 an hour, and for many it was the first honest look at the job.
When a tool takes the typing, it takes the seat in the corner too. No one decided to remove the apprenticeship. It went with the line item.
This is a small example of a larger pattern. Every documentation tool changes who is in the room and what they get to learn. We tend to measure the minutes saved on the note. We rarely measure what the note-taking used to carry.
Why it matters on the ward
Residents and juniors learn the same way, through proximity. They write the note, and writing it forces them to commit to a story about the patient. Dr. Taylor's concern about deskilling is about exactly that: if the draft arrives finished, who does the reasoning, and who learns to?
We do not have data on that from this story. The article is a student newspaper report from one health system, and it should be read that way. It is a scene, not a trial. But it is a scene worth taking seriously, because the incentives behind it are not unique to Virginia.
What a better design would keep
If documentation gets automated, the learning that rode along with it needs a deliberate home. A few questions worth asking of any rollout:
- Who used to learn from this task, and where do they learn it now?
- Does the junior still have to form their own assessment before the draft appears?
- Is the clinician reviewing the output as an author, or just signing it?
These are workflow questions, not model questions. They are also the reason we think the doctor's own workspace matters. A tool that helps a doctor think through a patient, rather than silently finishing the paperwork, keeps the reasoning with the person who needs to build it.
Sources
Medows is a clinical AI workspace for the doctor on rounds. Learn more or write to alapan@medows.ai / alapanx@gmail.com.