Structural ageism in big data approaches

Abstract

Digital systems can track every activity. Their logs are the fundamental raw material of intelligent systems in big data approaches. However, big data approaches mainly use predictions and correlations that often fail in the prediction of minorities or invisibilize collectives, causing discriminatory decisions. While this discrimination has been documented regarding, sex, race and sexual orientation, age has received less attention. A critical review of the academic literature confirms that structural ageism also shapes big data approaches. The article identifies some instances in which ageism is in operation either implicitly or explicitly. Concretely, biased samples and biased tools tend to exclude the habits, interests and values of older people from algorithms and studies, which contributes to reinforcing structural ageism.