Computing That Serves

End-User Control of Intelligent Systems: The Role of Interactivity in Computer-Supported Decision Making


Thursday, November 17, 2016 - 11:00am


Jacob Solomon


Kent Seamons

Jacob Solomon
Thursday, November 17, 2016
11:00am  1170 TMCB

Interaction with computers has become integrated into many of our decision-making processes, helping us decide which route to take on a commute, what book to read on a plane, when to call a doctor about a symptom, how to invest money, who to date, or what medication to prescribe to a patient. Human-centered computing research supports the idea that highly interactive interface designs that give users control and autonomy create a better user experience. But do interactivity and end-user control of intelligent information systems actually help people make better decisions? 
I report a series of studies that show how intelligent systems that afford to users control and customizability can trigger decision-making biases, most notably that users do a poor job of recognizing that a system is wrong when they have a lot of control over how it works. While this presents a challenge for designing effective decision support systems, there are also opportunities to leverage these biases to improve computer-supported decision making. I will describe an effort to do this in a tool I have built to help clinicians make better decisions about cancer screening.


Jacob Solomon is a Postdoctoral Fellow at the University of Michigan Medical School. He earned a PhD in Media and Information Studies in 2015 from Michigan State University, a Master's degree in Human-Computer Interaction in 2010 from the University of Michigan, and Bachelor's degrees in Psychology and Russian in 2007 from the University of Utah. His research combines social science inquiry with system building to study how socio-technical system designs affect user behavior and decision making. He has published work on the design of recommender systems, crowdfunding platforms, online communities, and instant messaging. His current work involves building and evaluating systems that support medical decision making by patients and healthcare providers.