Duke Pioneers New Approaches for Implementing and Evaluating AI in Healthcare

Duke University researchers are focusing on the safe and fair application of artificial intelligence (AI) in healthcare, particularly with Large Language Models (LLMs) like ChatGPT. These tools enhance administrative efficiency by automating tasks such as medical note-taking and responding to patient queries. To address concerns over ethical usage, the researchers emphasize the importance of AI governance for transparency and accountability. A specific innovation, Ambient Digital Scribing (ADS), helps reduce physician “pajama time” by transcribing consultations, allowing doctors to focus more on patient interaction and regain valuable time in their schedules. Michael Pencina, Duke’s Chief Data Scientist, leads efforts to ensure AI tools are rigorously evaluated before patient use, while colleague Chuan Hong is developing AI frameworks, like SCRIBE, for ongoing assessments of these technologies. This approach aims to enhance AI efficiency while minimizing human workload, fostering a responsible integration of AI in clinical settings.

Source link