Privacy and Fairness through Accountability
Anupam Datta
Abstract:
Organizations that collect and use personal information at scale should be held accountable for practices that violate societal values like privacy and fairness. This talk will provide an overview of the emerging research area of computational accountability -- theories and tools that can be used to provide oversight of complex information processing systems to detect and assign responsibility for violations of values like privacy, fairness, and transparency. I will cover two results in some detail: (1) ScopeScan: A methodology and tool chain for internal oversight of big data systems (developed jointly with Microsoft Research) that is now in use on Microsoft's production system, daily checking privacy compliance of Bing's big data analytics pipeline for compliance with its privacy policies. (2) AdFisher: A methodology and tool chain for external oversight of online services and its application to discover discrimination and opacity in the Google ad network. I will also cover some exciting open problems in this research area.
Bio:
Anupam Datta is an Associate Professor of Computer Science and Electrical & Computer Engineering at Carnegie Mellon University and a former occupant of Gates 468.