The Corey system functions through a flow of data input through both direct user interactions via cell phone, smart watch, and automatic input from peripheral devices with the user’s permission. Automatic input could range from heart rate, sleep, and exercise data gathered through health application, to the user’s grocery list collected through natural language processing technologies of voice-based home devices. The user-generated data, along with relevant medical literature, is used to create pertinent and timely suggestions and predictions that inform and initiate user action. Corey and the user may take additional actions that are prompted through the same series of devices. 

Reflections and Implications

Machine learning and healthcare are subjects that often feel sterile and intimidating, but with thoughtful design, machine learning can make healthcare knowledge approachable, and provide otherwise unforeseen connections that can help patients understand and navigate changes in their condition. Our research in the design of Corey highlights the important role delight and play have in making a user’s experience with complex and sensitive subjects approachable and impactful. Our findings also speak to the broader potential for machine learning’s capacity as an empathy-building tool.

Our designs for Corey rest on several assumptions. A user’s health data is extremely personal and sensitive information, therefore Corey is situated in a future in which patients own and control their own healthcare data via a single, secure, comprehensive access point.