A new randomized clinical trial conducted at UCLA Health shows that artificial intelligence (AI) tools designed to automatically document patient encounters can reduce the time doctors spend on paperwork and improve their work experience.
The study, published in the New England Journal of Medicine AI, investigated two commercially available AI scribe applications (Microsoft DAX and Nabla) in real-world clinical settings. In contact with 238 physicians and 72,000 patients across 14 specialties, researchers found that Nabla users reduced documentation time by nearly 10% compared to usual care (control group), while both tools showed potential benefits for physician burnout and work-related stress.
“Documentation burden is a major cause of physician burnout, with physicians often spending two hours documenting for every hour of patient care,” said lead author Dr. Paul Lukac, chief AI officer at UCLA Health. “This is the first randomized trial to rigorously assess whether AI scribes are delivering on their promise to help address this problem.”
Research structure
The research team randomly assigned doctors to use one of two AI writing tools or continue with their normal writing tasks for two months from November 2024 to January 2025. The AI scribe records patient conversations and automatically generates draft clinical notes for doctors to review and edit.
Physicians using Nabla saw the average time spent writing each note decrease by an estimated 41 seconds (from 4 minutes 30 seconds to 3 minutes 49 seconds), compared to a decrease of 18 seconds (from 4 minutes 22 seconds to 4 minutes 4 seconds) in the control arm. The decrease in the Nabla group was 9.5% greater than the control group, a statistically significant result. The DAX group showed a small decrease that did not reach statistical significance compared to the control group.
Importantly, both AI tools showed modest improvements in validated measures of physician burnout, cognitive workload, and job fatigue, but these findings need to be confirmed in larger studies. For example, physicians in the Nabla and DAX groups had approximately 7% improvement in burnout scores compared to physicians in the control group.
Balance benefits and concerns
This study also revealed important limitations. Physicians reported that AI-generated notes “occasionally” contained clinically significant inaccuracies, most commonly missing information and incorrect pronouns. One minor patient safety event was reported during the study.
“This technology requires active physician oversight, rather than passive acceptance,” said senior author Dr. John N. Mafi, an internist at UCLA Health. “Our trials revealed that while AI scribes offer measurable benefits, they can also create clinically significant inaccuracies. Physicians must remain vigilant when reviewing AI-generated documents. Moving forward, we must embrace innovation while maintaining medicine’s fundamental commitment to patient safety through rigorous evaluation and continuous monitoring.”
Survey responses indicated that physicians found both tools easy to learn and use and felt they were able to better engage with patients. Patients generally accepted the technology, with less than 10% refusing its use.
Significance for healthcare
Physician burnout affects nearly half of U.S. physicians and contributes to workforce shortages, increased medical errors, and billions of dollars in costs to the health care system. While electronic medical records have improved many aspects of care, they have significantly increased documentation demands.
This study provides timely evidence as health systems across the country are rapidly implementing AI scribes, whose effectiveness and safety are often not rigorously evaluated.
“By incorporating randomized trials into daily practice, we have been able to provide high-quality, real-world evidence to guide decision-making regarding the implementation of AI in healthcare,” said Lukac. “This approach serves as a model for responsibly evaluating other AI tools as they emerge.”
The researchers note that because the study was conducted at a single academic medical center over a relatively short period of time, the findings may not be broadly applicable to all clinical settings. They are calling for longitudinal studies across multiple sites to confirm these results and measure downstream effects on important health outcomes such as quality of care, cost, and experience.
Note: In the spirit of this research on AI-assisted clinical documentation, this press release was drafted with assistance from Claude (Anthropic) and reviewed and edited by UCLA Health Communications staff.
Research content
Article: Ambient AI Scribing in Clinical Practice – A Randomized Trial. NEJM AI 2025; DOI: 10.1056/AIoa2501000
Author and affiliation:
Paul J. Lukac, MD, MBA, MSc. William Turner, BS; Sitaram Vangara, MS. Aaron T. Chin, MD. Joshua Kalili, MD. Dr. Yachen Tina Shi. Eric M. Chen, MD, MSc; John N. Maffey, MD, MPH (David Geffen School of Medicine, University of California, Los Angeles)
Catherine Sarkisian, MD, MSHS (UCLA David Geffen School of Medicine, Los Angeles Department of Veterans Affairs)
John N. Maffi, MD, MPH is also affiliated with the RAND Corporation in Santa Monica, California.
Funding was provided by the UCLA School of Medicine, the UCLA Faculty Practice Group, and the National Institutes of Health/National Institute on Aging Beeson Emerging Leaders in Aging Research Career Development Award (grant K76AG064392-01A1). Support for the project was also provided by the UCLA Healthcare Value Analytics and Solutions (UVAS) consortium. The study protocol was registered at ClinicalTrials.gov (NCT06792890) and approved by the UCLA Institutional Review Board and Health AI Council.
/Open to the public. This material from the original organization/author may be of a contemporary nature and has been edited for clarity, style, and length. Mirage.News does not take any institutional position or position, and all views, positions, and conclusions expressed herein are those of the authors alone. Read the full text here.
