Stanford researchers have introduced ChatEHR, an experimental AI assistant that enables clinicians to query electronic health records using natural language. Powered by a version of GPT-3.5-turbo and integrated with Stanford’s Epic EHR system, this tool aims to reduce documentation time and improve clinical decision-making. In a controlled test at Stanford Hospital, physicians rated ChatEHR as helpful in 88 percent of interactions, particularly for retrieving lab results and generating patient summaries. Human oversight and safeguards are in place to monitor hallucinations and ensure data privacy.
The design emphasizes safety and trust through on-device processing, clinician review of responses, and strict oversight protocols. ChatEHR is part of a broader NIH-funded initiative to apply AI in healthcare without compromising patient safety or clinician accountability. While promising in reducing clinician burnout and improving efficiency, researchers stress the need for continued evaluation before deployment in real patient care settings.
Stanford Medicine. (2025, June 12). Stanford tests AI chatbot that helps doctors with electronic health records. https://med.stanford.edu/news/all-news/2025/06/chatehr.html
Comments
Post a Comment