What research says about AI documentation: less screen time, less burnout
AI that writes draft journal entries has gone from being a promise to being tested on a large scale. So, what does the research actually show – and what is it honest about that the technology doesn't solve alone?
In recent years, a number of studies have emerged on what is often called ambient documentation: tools that listen to consultations and create draft journal entries. The interest is linked to a well-known problem. The documentation burden is one of the biggest drivers of burnout among physicians, and an often-cited estimate is that physicians spend about two hours on paperwork for every hour with a patient. We have previously written about how Documentation can be the straw that breaks the camel's back for healthcare professionals, leading to burnout.. The question is whether technology actually helps – and research is now beginning to provide answers.
A randomized study: less time, lower burnout
Perhaps the most robust study to date is a randomized clinical trial conducted at UCLA Health and published in New England Journal of Medicine AI. According to UCLA Health The study included 238 physicians across 14 specialties and approximately 72,000 patient encounters, from November 2024 to January 2025. Physicians were randomly assigned to use ambient tools or continue as usual.
The findings are cautiously positive. One tool showed a statistically significant reduction in time spent per note compared to the control group. Even more interesting, both tools demonstrated a modest but measurable improvement in validated measures of burnout, cognitive workload, and fatigue – on the order of about 7 percent better burnout scores than the control group. Physicians also reported that the tools were easy to learn and that they experienced better patient engagement. Fewer than ten percent of patients opted out of using them.
The research is honest about the limitations
What makes this study worth taking seriously is that it doesn't hide its weaknesses. Doctors reported that the AI-generated notes «occasionally» contained clinically significant inaccuracies – most often omissions of information or pronoun errors – and one mild patient safety incident was reported during the study.
The researchers' conclusion is clear: the technology requires active oversight from the physician, not passive acceptance. An AI draft is just that, a draft, and its value depends on a human reviewing, correcting, and approving it. It's also worth remembering that the study was conducted at one academic hospital over a relatively short period, so the findings need to be confirmed in larger and longer studies across multiple settings.
This image is supported by broader knowledge summaries. A Systematic review from the Norwegian expert community for minimal method assessment compared ambient documentation with traditional methods in an outpatient setting, pointing to promising results while emphasizing the need for more and better documented evidence. Overall, the research paints a sober but optimistic picture: the technology can save time and ease the burden, provided it is used with control.
What does this mean in practice?
For the individual clinician, the message is both uplifting and sobering. Ambient documentation can free up time and alleviate some of the burden leading to burnout – but only if humans remain in control. This aligns well with how authorities are thinking: national guidelines for the safe use of AI in documentation emphasize that healthcare professionals are responsible for the content of the medical record and must quality-assure the draft before it is saved.
In other words, it’s not a question of «AI or human,» but of AI and Human: The technology takes the first draft, the clinician takes responsibility.
This is how Medivox is built around the same findings
Medivox is developed based on precisely this premise. The tool listens to the consultation and delivers a structured draft for a medical record note in Norwegian – but it is always you who reviews, corrects, and approves before anything is saved. You own the medical record and make the final assessment. It is not a limitation, but the whole point: research shows that the value lies in the combination of rapid draft generation and human control.
Because you build your own templates, the draft is customized to your practice and professional terminology, making it easier to capture those very details studies warn can be missed. And privacy is protected: personal data is pseudonymized before further data processing, and all data processing takes place in Norwegian data centers.
Frequently Asked Questions
Does AI documentation actually reduce time spent on charting?
Research shows promising but varied results. In a randomized study at UCLA Health, one ambient tool provided a statistically significant reduction in time per note compared to the control group.
Does that have an effect on burnout?
Yes, a modest effect. In the same study, the tools showed around 7 percent better scores on validated measures of burnout than the control group, but the researchers emphasize that the findings need to be confirmed in larger studies.
Can you blindly trust AI-generated notes?
No. The studies found that the drafts occasionally contained clinically significant inaccuracies. The conclusion is that the technology requires active supervision – humans must review and approve.
How does Medivox handle this?
Medivox creates a draft that you always review, correct, and approve. You own the record and make the final assessment, and custom templates make it easier to capture details.
Are the patient details safe?
Personal data is pseudonymized before it is processed further, and all data processing takes place at Norwegian data centers.
Use Medivox for free – Get started completely free
Curious how a documentation tool built for a combination of efficiency and control works in practice? Contact us – then we'll show you how Medivox works together with your professional knowledge.
Sources:
- UCLA Health (2025): UCLA study finds AI scribes may reduce documentation time and improve physician well-being
- Mini-method assessment (2026): A Systematic Review Comparing Ambient Scribes and Conventional Documentation Methods in Outpatient Care