Artificial Intelligence Scribe and Large Language Model Technology in Healthcare Documentation: Advantages, Limitations, and Recommendations
Problem
Physicians spend substantial time and cognitive effort on medical documentation during patient encounters.
Why it matters
Documentation burden reduces physicians' cognitive freedom and patient-facing time.
What the AI did
AI scribe applications using an LLM API (e.g., GPT-4) with automatic speech recognition generate full medical notes and draft follow-up e-mails within seconds to minutes.
Results
not stated in source
Impact
AI scribes offer efficiency gains in early adoption, but physician proofreading remains essential due to possible errors of omission, fabrication, or substitution.
Record
| Venue | Plastic & Reconstructive Surgery Global Open |
|---|---|
| Published | 2025 |
| Work type | article |
| Authors | Sarah A. Mess, Alison Mackey, David Yarowsky |
| Citations | 82 (OpenAlex, at import time) |
| References | 25 |
| Model or method | Large language model (generative pretrained transformer 4) with automatic speech recognition |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1097/gox.0000000000006450 |
| Open access | gold |
| Keywords | AI scribe, large language model, medical documentation, GPT-4, electronic medical records |
Importance rating
significant — thousands to millions affected
Abstract states AI scribes offer 'unprecedented efficiency' and cognitive freedom by cutting documentation time, affecting physicians broadly; no scale quantified and error/oversight risks noted.