Data quality dimensions for fair AI

TL;DR

This paper examines how data quality dimensions like completeness, consistency, timeliness, and reliability influence bias in AI systems, proposing formal methods to improve fairness especially in gender classification tasks involving non-binary and transgender individuals.

Contribution

It introduces a data quality framework for assessing bias in AI, extending fairness considerations beyond accuracy to include multiple data quality dimensions with formal reasoning.

Findings

Bias mitigation tools can be improved by considering data quality dimensions.

Formal methods help analyze the impact of data inconsistencies on fairness.

Addressing data quality can reduce bias in gender classification tasks.

Abstract

Artificial Intelligence (AI) systems are not intrinsically neutral and biases trickle in any type of technological tool. In particular when dealing with people, the impact of AI algorithms' technical errors originating with mislabeled data is undeniable. As they feed wrong and discriminatory classifications, these systems are not systematically guarded against bias. In this article we consider the problem of bias in AI systems from the point of view of data quality dimensions. We highlight the limited model construction of bias mitigation tools based on accuracy strategy, illustrating potential improvements of a specific tool in gender classification errors occurring in two typically difficult contexts: the classification of non-binary individuals, for which the label set becomes incomplete with respect to the dataset; and the classification of transgender individuals, for which the dataset becomes inconsistent with respect to the label set. Using formal methods for reasoning about the behavior of the classification system in presence of a changing world, we propose to reconsider the fairness of the classification task in terms of completeness, consistency, timeliness and reliability, and offer some theoretical results.