CBDS Digital Health Seminar: “Beyond algorithmic bias: AI decision-support systems practitioners can trust and use effectively” by Dr. Rene Kizilcec, Cornell University

Data-driven algorithms can offer insights that improve the judgements and decisions of both novices and experts, but their benefits are contingent on successfully integrating these systems in existing practices. The adoption and effective use of an AI decision-support system requires that users trust it and comprehend how it works, and especially how it can complement them. Extensive research on algorithmic bias, while important for not amplifying inequities, fails to consider how an AI’s bias compares to the status quo of human decisions and it can erode trust in algorithms for people who would be better off using them. Effectively integrating algorithms in clinical and educational practices so that they are accepted, used effectively, and improve outcomes is just as important, I argue, as developing such algorithms and auditing them for bias. In this talk, I will cover studies of human factors in AI decision-support systems and their implications for how to design and integrate algorithms in practice.

Rene Kizilcec is an Assistant Professor of Information Science at Cornell University, where he directs the Future of Learning Lab. He studies the use and impact of technology in formal and informal learning environments (incl. college classes, online degree programs, mobile learning, professional development, MOOCs, and middle/high school classrooms) with behavioral, psychological, and computational methods. His research on algorithmic fairness and transparency examines equity-based concerns about predictive models in the real world. His work on scalable interventions to broaden participation and reduce performance gaps appeared in Science and PNAS. Kizilcec received a B.A. in Philosophy and Economics from University College London, and a M.Sc. in Statistics and Ph.D. in Communication from Stanford University.