How to Trust AI in Healthcare: MLOps

Monday, August 9 at 11:55 AM - 12:10 PM PDT
Wynn, Bandol 2
 AI models are increasingly prevalent in healthcare and hold the promise of dramatically improving patient outcomes and operational efficiency. However, deployed AI models require extreme care to avoid dangerous extrapolation and maintain accuracy. Healthcare Organizations should demand that every deployed AI model is monitored for data drift, accuracy, and bias / fairness. They should utilize automated alerts, concrete business or clinical rules, and a champion / challenger deployment strategy to mitigate the risks of wayward AI models


Sally Embrey
Vice President, Public Health and Health Technologies