Abstract
Algorithmic fairness is just one piece of a larger goal of using computing and data science to address problems in human welfare and social justice. Here at UC Berkeley, we have recently founded the Division of Computing, Data Science, and Society, which comprises EECS, Statistics, the School of Information, and interdisciplinary partnerships across the entire university and with UCSF. One of the primary research themes of our new division is computing and data science for human welfare and social justice. We plan to partner with the UC Berkeley School of Public Health, School of Social Welfare, Graduate School of Education, Law School, and School of Public Policy to develop a broad view of human welfare both internationally and in this country. In particular, in the US we plan to look at data from our public health system, social welfare system, K – 12 educational system, and criminal justice system, and to work with UC Berkeley alumni in all of these fields, to frame the problems. What inferences can we draw from this heterogeneous data, and how can we use these inferences to suggest effective policy and interventions? What new data should we be trying to collect? Which parts of these systems are becoming algorithmically driven, and how do we make sure these algorithms are as fair as possible?
I will speak for about 10 minutes about what we are doing at Berkeley. Then I will facilitate a panel including Rediet Abebe, Junior Fellow, Society of Fellows, Harvard; Linda Burton, Dean and Professor, School of Social Welfare, UC Berkeley; John Eason, Professor of Sociology, University of Wisconsin; Marzyeh Ghassemi, Assistant Professor of Computer Science and Medicine, U. Toronto; UC Berkeley; and Ziad Obermeyer, Associate Professor of Health Policy Management, School of Public Health, UC Berkeley.