The Learning Curve: What AI Scribes Mean for the Next Generation of Clinicians
New research from Yale asks what happens when AI scribes enter medical training — and every teaching practice needs a thoughtful answer.
Most of the ambient AI scribe conversation assumes an experienced clinician on the other end of the microphone — someone who already knows how to construct a differential, who reads the AI draft with a trained eye, and who catches what the machine misses. But a growing body of work is asking a harder question: what happens when the person using the scribe is still learning to think and document like a physician?
New research from Yale School of Medicine examining the impact of AI scribes on student note quality has pushed this issue into the open, and it deserves the attention of every teaching practice, residency program, and clinical educator. The Yale study assessed how an ambient AI scribe affected the quality of clinical notes written by first-year medical students during a structured clinical examination — including “hybrid” notes that merged AI-generated drafts with student writing. Documentation isn’t clerical busywork. Writing the note is where a trainee organizes a chaotic encounter into a reasoned clinical narrative. If AI writes that note, does the learning still happen?
Do AI Scribes Undermine Clinical Reasoning Skills?
The short answer from educators is: it depends entirely on how the tool is used. Passive acceptance of AI drafts risks eroding the reasoning that documentation is meant to build, while active critique of those drafts can sharpen it. The Yale work matters because it targets the earliest, most formative stage of training — first-year students — where documentation habits are still being set. What happens at that stage tends to persist for a career.
Documentation as a Cognitive Skill
Ask any senior physician how they learned to reason clinically, and many will point to the note. Artificial intelligence (AI) can now generate that note for them — which is precisely what makes the question urgent. The act of writing forces synthesis: you must decide what’s relevant, sequence a history, commit to an assessment, and justify a plan. It is deliberate practice in written form. A trainee who struggles to write a coherent assessment is often a trainee who hasn’t yet fully reasoned through the case — and the writing exposes the gap so it can be addressed.
An ambient scribe that hands a learner a polished, pre-reasoned note short-circuits this loop. The trainee edits rather than constructs. The struggle that builds skill is quietly removed, and with it, potentially, some of the learning.
The Double-Edged Draft
This isn’t an argument against AI in training environments — the technology is coming whether or not educators are ready, and it carries genuine benefits. Used well, an AI draft can be a teaching artifact: a starting point a preceptor and trainee critique together, spotting what the AI over-documented, what it missed, and where its clinical framing was thin. In that mode, the scribe becomes a shared object for feedback rather than a shortcut.
The risk is the passive mode — where the trainee accepts the draft, learns to trust it, and never develops the independent muscle. The difference between these outcomes is not the tool. It’s the pedagogy wrapped around it.
Guidance for Teaching Practices
Programs integrating ambient documentation into training environments can protect learning while embracing the technology:
- Scaffold, don’t skip. Have junior trainees draft notes independently first, then compare against the AI output. The comparison itself is the lesson.
- Make the AI critiqueable. Teach learners to interrogate AI drafts — to find omissions, weak assessments, and misplaced emphasis — as a core competency, not an afterthought.
- Stage autonomy. Reserve heavier AI reliance for later-stage trainees who have already demonstrated documentation competence independently.
- Assess reasoning, not just output. If notes look uniformly polished, shift evaluation toward oral case presentations and the trainee’s ability to defend their clinical thinking.
Takeaway
Ambient AI scribes will be part of medicine for the entire careers of today’s trainees, and pretending otherwise does students no favors. But documentation has always been more than paperwork — it’s where clinical reasoning is forged. Teaching practices that treat the AI draft as a subject of critique rather than a substitute for effort will graduate clinicians who can both use the tool and see past it. That is the skill that will matter most.
MyMediScribe supports clinician-in-the-loop workflows that keep human judgment — and human learning — at the center of the note.
See how MyMediScribe keeps your documentation clinician-first at mymediscribe.com/security.