Despite the volume of articles about the potential for AI to improve patient outcomes and health care operations, finding information about how to responsibly deploy machine learning models in hospitals remains difficult. Only a tiny fraction of health AI research finds its way into the real world, and current deployment strategies are ad hoc, lacking defined rules and best practices. To better showcase examples of AI moving from the lab to the clinic, the Vector Institute and the Hospital for Sick Children (SickKids) organized the Vector-SickKids Health AI Deployment Symposium in Toronto on October 30, 2019. 

In providing examples of successfully developed and deployed systems, the new white paper coming out of the symposium, “Implementing AI in Healthcare,” aims to initiate a discussion around the key ingredients for success that will enable AI and machine learning developments to have a measurable and positive presence at the bedside. Coming from across North America, speakers shared their successes and challenges, providing valuable insights into this new and evolving field.