Peer-reviewed research
A Pragmatic Randomized Controlled Trial of Ambient Artificial Intelligence to Improve Health Practitioner Well-Being
Afshar M et al., NEJM AI, 2025 · DOI 10.1056/aioa2500945
Read the originalRead on October 16, 2026
- What it is
- A stepped-wedge randomized trial at the University of Wisconsin Hospitals and Clinics. Sixty-six practitioners were randomly assigned to three 6-week sequences of ambient AI, and the trial measured professional fulfillment, work exhaustion, time spent on notes and documentation quality.
- How much weight it carries
- Peer-reviewed, with a randomized design and 71,487 notes in the analysis. It covers one health system and 66 practitioners. The abstract reports a P value of .04 for professional fulfillment but calls that change nonsignificant, so the full paper should be read before quoting that result.
Facts taken from this source
- The University of Wisconsin trial randomly assigned 66 health care practitioners to three 6-week sequences of ambient AI.
“Sixty-six health care practitioners were randomly assigned to three 6-week sequences of ambient AI.” (Methods)
- In the Wisconsin randomized trial, time spent on notes fell by 0.36 hours per day with ambient AI use, a secondary measure.
“time spent on notes decreased (-0.36 hours per day; 95% CI, -0.55 to -0.17)” (Results)
- In the Wisconsin randomized trial, ambient AI use reduced work exhaustion and interpersonal disengagement by 0.44 points on a five-point scale.
“Ambient AI use had a significant reduction in work exhaustion/interpersonal disengagement (-0.44 points; 95% confidence interval” (Results)