ai systems are not intrinsically neutral and biases trickle in any type of technological tool. In particular when dealing with people, AI algorithms reflect technical errors originating with mislabeled data. As they feed wrong and discriminatory classifications, perpetuating structural