Preventing harm from non-conscious bias in medical generative AI (December 2023)

Summary

Large language models such as OpenAI's GPT-4 have the potential to transform medicine by enabling automation of a range of tasks, including writing discharge summaries, answering patient questions, and supporting clinical treatment planning. These models are so useful that their adoption has been immediate, and efforts are already well underway to integrate them with ubiquitous clinical information systems. However, the unchecked use of the technology has the potential to cause harm. In this article for The Lancet, Janna Hastings looks at the need to mitigate racial and gender bias in language models that may be used in healthcare settings.