AI Discharge Summaries Win Physician Preference
A Stony Brook University study found physicians preferred AI-generated hospital discharge summaries over doctor-written ones in most cases.
Updated on Nov. 7, 2026 in Artificial Intelligence
Live Poll
Would you trust an AI-written summary of your hospital discharge instructions?
Researchers at Stony Brook University published a study in October in the journal Digital Medicine showing that AI-generated posthospital discharge summaries outperformed those written by doctors. Physician reviewers preferred the AI summaries in 95% of 60 cases.
Why it matters
Discharge summaries guide the posthospital care team, and writing them can take 15 minutes or more than an hour. The authors said reducing clinician cognitive burden and improving transitions of care motivated the work.
The study compared summaries for 60 cases using GPT-5.2, an OpenAI model released in December 2025, and reviewers found no significant difference in estimated harm potential between AI and doctor-written summaries. Whether these results hold in prospective, real-world clinical use remains untested.
The players
Stony Brook University
A public research university in New York whose researchers conducted the study comparing AI and physician-written discharge summaries.
OpenAI
An artificial intelligence company whose GPT-5.2 model, released in December 2025, generated the summaries evaluated in the study.
Dr. Morgan
A physician at the University of Maryland School of Medicine who commented on the study's findings.
The details
Physician reviewers judged the summaries on ease of understanding, communication to patients and caregivers, and usefulness for follow-up care. The study used existing clinical data, and doctors spend about two or three times as much time in front of a computer screen as talking to patients.
Timeline
December 2025: OpenAI released GPT-5.2, which was state-of-the-art at release.
October 2026: The Stony Brook study was published online in Digital Medicine.
Next six to 18 months: Commercial applications of AI summarization tools are possible.
Big Picture
The study follows the pattern set by general-purpose language models such as OpenAI's GPT-series being evaluated on clinical documentation tasks. It tests whether that pattern holds for the posthospital discharge summary, a document physicians have long written by hand or by dictation.
If adopted, AI-drafted discharge summaries could reduce the time physicians spend on documentation and affect the information patients and caregivers receive after leaving the hospital. Hospitals would likely pilot the tools before using them routinely.
The takeaway
Clinicians and patients may benefit if AI-drafted summaries are validated in real-world settings, since the study's results were retrospective. Readers should watch for prospective trials before expecting changes to their own discharge paperwork.
Further reading
Explore more on Artificial Intelligence developments shaping medicine and clinical workflows.
Source note: This article includes information reported by Newsday.
Live Poll
Would you trust an AI-written summary of your hospital discharge instructions?
