Abstract

Information access systems, such as search engines, recommender systems, and conversational agents, are used daily by billions of Internet users and have a profound impact on users’ information experiences, access to knowledge, and understanding of the world and people around them. These systems differ in crucial ways from the kinds of systems most frequently studied in the algorithmic fairness literature, requiring new techniques to properly understand and measure their social impacts. In this talk, I will discuss what makes these systems different and interesting; ground the quest for fairness and mitigating social harms in relevant ethical and legal considerations; and describe key considerations in measuring and mitigating harms that also apply beyond information access to a wider range of computing systems.