Racist algorithms in Google Image Search

a critical platform discourse analysis

Abstract

Debates around ethno-racial issues have gained increasing momentum in contemporary society, as they problematize and challenge behaviors rooted in colonial logics. This is largely due to the efforts of anti-racist Black movements that operate daily in both online and offline spaces. In this context, the present study aims to reflect on and interrogate the algorithmic practices of Google Image Search that perpetuate anti-Black racist discourse. To that end, we draw on the interdisciplinary principles of Applied Linguistics, which enable dialogue with theories from other fields, such as Platform Studies (D’Andréa, 2020) and the concept of Antiblackness (Vargas, 2020). Methodologically, this is a qualitative, exploratory case study focused on Google’s image bank, given its accessibility and widespread use across different age groups. The searches were divided into two categories — aesthetic and economic — based on comparative and individual search commands such as “successful people vs. poor people,” “beautiful women,” and “woman with beautiful hair.” The results indicate that the discourses revealed through these searches contribute to the perpetuation of racist stereotypes, as algorithms are not neutral. Rather, they enable the reproduction of harmful social practices that were once confined to analog environments, now amplified in digital spaces through the internet’s wide reach.