AI in Healthcare

How will medicine change in the world of AI?

We believe that medicine needs to understand the needs of patients and communities, and that these are not adequately included in current AI systems. We’re building tools to change that.

  • How can we evaluate bias in AI systems? Can we test directly for differences in care recommendations? We are developing a web app to graph AI medical assistant responses by race, sex, and age to visualize bias in AI recommendations.

  • AI systems are being rapidly deployed in healthcare without sufficient testing. We are developing methods to include community feedback in the training of AI models so that marginalized voices can be part of AI-powered healthcare.

Collaborators

  • T Templin

    UNC Chapel Hill

  • Jeff Leek

    Fred Hutchinson Cancer Center