Unequal Representations: Analyzing Intersectional Biases in Word Embeddings Using Representational Similarity Analysis
This research highlights the presence of intersectional biases against Black women in word embeddings, which is a problem for fairness in AI systems.
This paper analyzes intersectional biases in word embeddings, revealing that Black women are represented as less feminine than White women and less Black than Black men. This finding supports intersectionality theory regarding unique modes of discrimination.
We present a new approach for detecting human-like social biases in word embeddings using representational similarity analysis. Specifically, we probe contextualized and non-contextualized embeddings for evidence of intersectional biases against Black women. We show that these embeddings represent Black women as simultaneously less feminine than White women, and less Black than Black men. This finding aligns with intersectionality theory, which argues that multiple identity categories (such as race or sex) layer on top of each other in order to create unique modes of discrimination that are not shared by any individual category.