ABSTRACT

Black women are vulnerable to discrimination not merely by virtue of being Black women, but because the law's single-axis thinking explicitly produces vulnerabilities for those who, like Black women, are multiply-oppressed. Today, it is widely recognized that data and algorithms risk reproducing biases against historically disadvantaged populations in ways that, as Barocas and Selbst put it, ‘look a lot like discrimination'. In the United States, this risk and efforts to mitigate it have in many ways echoed liberal antidiscrimination discourses in the law, which have historically sought to address injustices in the distribution and exercise of important rights, opportunities, and resources in domains like voting, housing, and employment. Single-axis thinking also tends to focus on relative disadvantage at the expense of attention to the production of systematic benefits or privileges. The last limitation centers on the tendency of antidiscrimination discourses to focus on disadvantage relative to a narrow set of goods, namely rights, opportunities, and resources.