If we do not fix medical documentation, your doctor keeps spending part of the visit being a stenographer with a medical license. Nobody went to school for that. Tierney and Lee’s Annals of Internal Medicine editorial, “Redefining Documentation Quality in the Age of Ambient Artificial Intelligence Scribes,” lands right on that sore spot and asks a blunt question: if AI is going to draft the chart, what counts as a good note now? [1]

Ambient AI scribes are the new favorite answer to the old EHR headache. They listen to the visit, use speech recognition to turn audio into text, try to separate who said what, then hand the transcript to a large language model that formats a clinical note, often in a SOAP-style structure. In plain English, it is a robot trying to sit quietly in the exam room and take decent notes. Quietly is doing a lot of work there.

The timing of Tierney and Lee’s editorial matters because it accompanies a much less dreamy paper. In a 2026 Annals evaluation, researchers compared notes from 11 ambient AI scribe tools with notes written by 18 clinicians across five standardized primary care scenarios. Human notes scored higher in every case. The biggest gaps showed up in thoroughness, organization, and usefulness, which are exactly the parts you would prefer not to get weird in a medical record. In the low back pain case with background noise, human notes averaged 43.8 out of 50 versus 20.3 for AI. That is not “close enough.” That is “please do not sign this without reading it twice.” [2]

Speed is nice. Missing details are less nice.

This is the core point. A clinical note is not just paperwork with a pulse. It is memory, communication, legal record, billing artifact, and future handoff all mashed into one mildly cursed document. If the note is sloppy, the next clinician inherits the sloppiness. If it misses nuance, the patient gets a cleaner paragraph and a worse outcome. Very efficient. Very modern.

That does not mean the whole idea is nonsense. Quite a bit of recent research says ambient scribes can reduce documentation burden. Stanford researchers reported reduced documentation and EHR time in an early multi-specialty pilot. [3] A JAMA Network Open study from Penn found clinicians using ambient scribing reported lower mental burden and better engagement with patients. [4] Another multicenter study linked ambient AI scribe use with reduced burnout, lower cognitive task load, and less time spent documenting. [5]

So yes, the tools can save time. The overworked interns doing all the actual math may, in fact, be helping.

But saved time is not the same as saved judgment.

What Tierney and Lee are really poking at

The title gives the game away. “Documentation quality” cannot mean only “the note appeared quickly and looked tidy.” A tidy lie is still a lie. A polished omission is still an omission. The whole field is now colliding with an awkward reality: AI scribes may improve workflow before they consistently improve documentation itself.

That tension shows up across the literature. A 2024 simulation study in Future Healthcare Journal found an ambient AI tool produced higher-quality documentation and shortened consultations by 26.3% on average. [6] A 2025 systematic review found mixed quality results overall, with some studies showing gains and others raising concerns about hallucinated or incomplete content. [7] In other words, the evidence is not a clean victory march. It is a pile of promising pilots, conflicting metrics, and a lot of “more study is needed,” which is science’s way of saying everyone should calm down a little.

Meanwhile, adoption is moving fast anyway. The Peterson Health Technology Institute wrote in March 2025 that ambient scribes were poised to become one of the fastest technology adoptions in healthcare history, with roughly 60 being implemented in practice. [8] NHS England has since issued guidance for deploying these systems, updated on March 30, 2026, which tells you the governance people have already started sweating. [9]

Fair enough. If software is listening to clinical visits, drafting records, and maybe nudging coding and reimbursement, “does it save five minutes?” is not the only question. It might not even crack the top five.

The boring answer is the correct one

Ambient AI scribes probably belong in medicine. Just not as magical replacements for human documentation. Not yet. Right now the strongest case is that they are draft tools. Useful draft tools, sometimes impressively useful, but still draft tools. The clinician remains the adult in the room.

That is why Tierney and Lee’s framing matters. The real standard for documentation quality should include accuracy, completeness, usefulness, patient context, and whether the note helps the next person make a safe decision. Speed gets a vote. It does not get the whole election.

Medicine has enough problems already. It does not need a note-taking assistant with the confidence of a man explaining crypto at a cookout.

References

1. Tierney AA, Lee K. Redefining Documentation Quality in the Age of Ambient Artificial Intelligence Scribes. Annals of Internal Medicine. 2026. DOI: 10.7326/ANNALS-26-00231

2. Reddy A, et al. Rapid Evaluation of Artificial Intelligence Technology Used for Ambient Dictation in Primary Care: Comparing the Quality of Documentation of Artificial Intelligence-Generated and Human-Produced Clinical Notes. Annals of Internal Medicine. 2026. DOI: 10.7326/ANNALS-25-02772

3. Ma SP, Liang AS, Shah SJ, et al. Ambient artificial intelligence scribes: utilization and impact on documentation time. Journal of the American Medical Informatics Association. 2025;32(2):381-385. DOI: 10.1093/jamia/ocae304. PubMed: 39688515. PMCID: PMC11756633

4. Duggan MJ, Gervase J, Schoenbaum A, et al. Clinician Experiences With Ambient Scribe Technology to Assist With Documentation Burden and Efficiency. JAMA Network Open. 2025;8(2):e2460637. DOI: 10.1001/jamanetworkopen.2024.60637. PubMed: 39969880

5. Olson KD, Meeker D, Troup M, et al. Use of Ambient AI Scribes to Reduce Administrative Burden and Professional Burnout. JAMA Network Open. 2025;8(10):e2534976. DOI: 10.1001/jamanetworkopen.2025.34976

6. Balloch J, Sridharan S, Oldham G, et al. Use of an ambient artificial intelligence tool to improve quality of clinical documentation. Future Healthcare Journal. 2024;11:100157. DOI: 10.1016/j.fhj.2024.100157. PMCID: PMC11452835

7. Bracken A, Reilly C, Feeley A, et al. Artificial Intelligence (AI)-Powered Documentation Systems in Healthcare: A Systematic Review. Journal of Medical Systems. 2025. DOI: 10.1007/s10916-025-02157-4

8. Peterson Health Technology Institute. Adoption of Artificial Intelligence in Healthcare Delivery Systems: Early Applications and Impacts. March 2025. Report

9. NHS England. Guidance on the use of AI-enabled ambient scribing products in health and care settings. First published April 27, 2025; updated March 30, 2026. Guidance

Disclaimer: This blog post is a simplified summary of published research for educational purposes. The accompanying illustration is artistic and does not depict actual model architectures, data, or experimental results. Always refer to the original paper for technical details.