Skip to content

National Library of Medicine explores using AI to respond to consumer health questions

Researchers at the National Library of Medicine are expanding a project to determine the role computers may play in interpreting and dispensing medical advice using artificial intelligence.

The project, which aims to develop intelligent computerized medical assistants for doctors, originally started with 9,000 medical questions from Stanford University and the University of Minnesota. However, researchers soon learned their efforts required a larger sample size. So they're partnering with Canada-based, a free, online medical information site. According to a release:

[Researchers] are using thousands of real-language, unedited medical questions at the National Institutes of Health (NIH) in Bethesda, Maryland to see if computers can interpret the tone and meaning of questions phrased by online patients. The NIH has also been in contact with IBM, discussing the possibility of working with IBM's Super Computer 'Watson' (of Jeopardy fame) to test these questions from

Milton Corn, National Institute of Health Deputy Director, noted in the release how users' questions were particularly valuable to the computational research project because they authentically reflected the language commonly used by patients.

Previously: Foundation launches contest to develop AI physician

Popular posts