In this text, I argue that the psychological study of implicit bias can benefit from second-person approaches. Specifically, I show that dominant experimental approaches based on the Implicit Association Test (IAT) presuppose an internalist account of implicit bias, according to which the possession of an implicit bias is identified with the possession of a conceptual association. By equating the possession of implicit bias with the possession of a conceptual association, the traditional view reduces bias to a particular activation in memory and ignores the sociocultural aspects involved in this phenomenon. To avoid this, I suggest adopting an externalist perspective, considering a second-person approach that allows sociocultural and normative aspects to be included in the identification of bias. Finally, I explore some consequences of the externalist proposal for the measurement and intervention on implicit bias.