Analyzing Epistemic Injustice

Cases of bias and unfair decisions in automated decision-making are heavily discussed. When unfair decision outcomes can be attributed to an unjust difference in the knowledge of various groups of subjects, we can speak of epistemic injustice (Fricker). In this paper, we analyze various kinds of epistemic injustice, such as testimonial, hermeneutical, distributional, and content-based epistemic injustice, and show how they can be conceptualized. We then provide a formalization of the difference in group knowledge, in a version of epistemic logic. After that, we discuss a case of a badly designed information system for government decision-making: Toeslagenffaire (Netherlands). We analyze key observations from the case and show that they constitute a form of epistemic injustice.