AI Mistakes Can Cost Doctors Time When Writing to Patients | Dartmouth Skip to main content Search<br>Search<br>search

More news

Articles<br>AI Mistakes Can Cost Doctors Time When Writing to Patients

News subtitle<br>Errors, extra details in AI-drafted responses can saddle physicians with more editing.

Image<br>Image

Computer science professor Sarah Preum and fellow researchers conducted the first large-scale study of an online patient portal that uses AI to draft responses from physicians to patients. (Photo by Katie Lenhart)

7/7/2026

More Reading<br>Dartmouth Researchers Assess Agentic AI

Body<br>Artificial intelligence is spreading rapidly in healthcare, with the goal of streamlining critical but onerous clerical tasks such as note-taking and charting so that physicians and nurses can devote more time to patients.<br>But even when AI can free up doctors to correspond with patients, it may fall short in helping them do it by introducing errors and extraneous details into their messages, according to a new Dartmouth study presented July 7 at the annual meeting of the Association for Computational Linguistics and published in the conference proceedings.<br>The result is that physicians may spend more time editing responses than it would take to write them, the researchers report.<br>“We find that AI can sound like a doctor but not think like one,” says Sarah Preum, an assistant professor of computer science and the study’s co-corresponding author with Parker Seegmiller, a Guarini School of Graduate and Advanced Studies PhD candidate in Preum’s PersistLab.<br>The researchers conducted the first large-scale study of an online patient portal that uses AI to draft responses from physicians to patients. The team developed a tool that compares AI-generated replies to a dataset of real responses they developed with healthcare professionals from Dartmouth Health.<br>They then analyzed 146,000 conversations between 10,105 patients and their primary care physicians at the large rural health system. The study was approved by the DH Institutional Review Board, and the team used the required methods to protect patient privacy, including anonymizing data as needed.<br>Quote<br>We find that AI can sound like a doctor but not think like one.

Attribution<br>Sarah Preum, assistant professor of computer science

The researchers also used their tool to evaluate physician responses drafted by Claude, Gemini, and ChatGPT, as well as the three smaller commercial platforms, Llama, Aloe, and Qwen.<br>The team reports that AI-generated answers frequently misalign with what clinicians would actually write. This includes automated responses that are too long, don’t ask follow-up questions, and use irrelevant or inaccurate medical details.<br>“There are smaller studies that say, ‘Oh, AI is amazing,’ but we realized there is a gap in the existing literature of a large-scale evaluation of this technology,” Preum says. “We didn’t just want to measure a platform’s accuracy, but whether it actually helps with the workload, which in this case is measured by how much editing the physician is doing.”<br>For example, the portal’s AI suggested telling a 32-year-old woman who is taking an acid reflux drug and was concerned about constant nausea that the medication might take some adjustment in diet. A physician replaced that by asking if there’s any chance she was pregnant.<br>Even little changes can add up over hundreds or thousands of messages, Preum says. “You don’t want to integrate large language models into the workflow and just shift the bottleneck so that doctors are devoting their cognitive energy to playing AI janitor and fixing mistakes,” Preum says. “But if we’re not careful, that’s a likely outcome.”<br>The researchers show, however, that adapting AI to how individual physicians communicate can improve accuracy by 33% and reduce editing by 26%.<br>“If message generation is really efficient and high quality, if it asks the right things, then it really has potential to improve efficiency,” says co-author Tim Burdick ’89, MED ’02, an associate professor of community and family medicine in the Geisel School of Medicine and a family medicine physician at DH.<br>“I don’t foresee a time when the portal can respond to a patient without a clinician editing it first. But as we make the models better, we’ll be able to address portal messages much more quickly and with less mental energy,” he says.<br>Burdick, Preum, and Seegmiller worked with co-authors Joseph Gatto, Guarini ’26; Sarah Greer, MED ’02, a former physician at Dartmouth Health; Ganza Belise Isingizwe ’26; and Rohan Ray ’26.<br>Image

Geisel professor Tim Burdick ’89, MED ’02, left, and Guarini PhD candidate Parker Seegmiller, helped lead the study.

The study shows that there are such things as “good” AI responses and provides a framework for implementing them into patient-physician portals, Preum says. These platforms are increasingly common among large healthcare systems and often customized, she says.<br>“That took us a long time to figure...