Skip to content

Work in progress

When machines discriminate. Artificial intelligence, gender bias and the labor market

Drawing on a benchmark experiment which compares popular supervised learning algorithms, this article investigates how the use of artificial intelligence by organisations induces gender discrimination in the hiring process. Algorithms reveal gender bias against women when predicting job applicants’ future success when hired. My findings highlight that motherhood penalty, a mechanism already well established in the labor market literature, plays a key role in explaining this result. Using panel data on the Italian labor market and performing conventional computations that resemble employers’ actual use of artificial intelligence, my findings contribute new understandings of gender discrimination in the application of automated decision making to the labor market, fostering the dialogue between sociology and computer science.