Author: Joy Buolamwini, Latanya Sweeney, and Darren Walker
Publisher: Radcliffe Institute for Advanced Study, Harvard University
Publication Year: 2019
Summary: The following video documents a conversation between Joy Buolamwini, Latanya Sweeney, and Darren Walker. Leading the discussion is Latanya Sweeney, founder of Public Interest Tech Lab. During this half-hour discussion these 3 subject matter professionals delve into many implications and some examples of algorithmic bias. Sweeney speaks from personal experience and describes how simple checks like modifying or diversifying the training data set completely changed the outcomes of her particular model. They spend several minutes talking about identification and facial recognition algorithms and how they tend to misattribute Black and Brown features at a much higher rate than other racial groups. They continue and identify the compounding effect for those effects groups as technologies become more advanced and integrated into everyday life. They purpose small scale-constant checks for biases well before any system is put in front of users. They also identify groups to become involved with like the Algorithmic Justice league.
