Interventions Against Machine-Assisted Statistical Discrimination
This addresses discrimination in algorithmic decision-making, offering a novel intervention beyond traditional methods like affirmative action.
The paper tackles statistical discrimination driven by machine learning by proposing a belief-contingent intervention called common identity, which effectively eliminates equilibrium discrimination even with biased training data.
I study statistical discrimination driven by verifiable beliefs, such as those generated by machine learning, rather than by humans. When beliefs are verifiable, interventions against statistical discrimination can move beyond simple, belief-free designs like affirmative action, to more sophisticated ones, that constrain decision makers based on what they are thinking. I design a belief-contingent intervention I call common identity. I show that it is effective at eliminating equilibrium statistical discrimination, even when training data exhibit the various statistical biases that often plague algorithmic decision problems.