Peer-reviewed research

Ambient AI Scribes in Clinical Practice: A Randomized Trial

Lukac PJ et al., NEJM AI, 2025 · DOI 10.1056/aioa2501000

Read the original

Read on October 16, 2026

What it is: A three-arm pragmatic randomized trial at UCLA that assigned 238 outpatient physicians to Microsoft DAX Copilot, to Nabla, or to usual care, and measured change in time spent writing notes from November 2024 to January 2025.
How much weight it carries: Peer-reviewed and randomized, with a published abstract that reports confidence intervals. It was run at one academic centre over about two months, and the vendors' products may have changed since, so the effect sizes may not transfer to other settings.

Facts taken from this source

The UCLA ambient scribe trial assigned 238 outpatient physicians across 14 specialties to two AI scribes or to usual care.
“238 outpatient physicians, representing 14 specialties, were assigned 1:1:1” (Methods)
The UCLA randomized trial of two AI scribes concluded that both products resulted in potential improvements in burnout, task load and work exhaustion, but that these secondary findings need confirmation in larger, multicenter trials.
“Both DAX and Nabla resulted in potential improvements in burnout, task load, and work exhaustion, but these secondary end point findings need confirmation in larger, multicenter trials.” (Conclusions)
In the UCLA randomized trial of two AI scribes, physicians rated clinically significant inaccuracies as occurring occasionally, for both products, on a five-point scale.
“clinically significant inaccuracies were noted "occasionally" on five-point Likert questions” (Results)
Physicians who used Microsoft DAX Copilot showed no statistically significant change in time-in-note compared with the control group.
“whereas DAX users exhibited no significant change versus the control group (-1.7%; 95% CI, -9.4% to +5.9%; P=0.66)” (Results)
In the UCLA randomized trial, physicians said clinically significant inaccuracies in the AI scribe notes occurred occasionally.
“clinically significant inaccuracies were noted "occasionally" on five-point Likert questions” (Results)
Physicians who used Nabla cut their time-in-note by 9.5 percent compared with the control group, the only arm with a statistically significant reduction.
“Nabla users experienced a 9.5% (95% confidence interval [CI], -17.2% to -1.8%; P=0.02) decrease in time-in-note versus the control group” (Results)

Ambient AI Scribes in Clinical Practice: A Randomized Trial

Facts taken from this source

Pages that rest on this source

Ambient AI Scribes in Clinical Practice: A Randomized Trial

Facts taken from this source

Pages that rest on this source