Intersectional Unfairness Discovery

TL;DR

This paper introduces BGGN, a generative network that efficiently discovers diverse high-bias intersectional sensitive attributes in AI systems, revealing potential unfairness in modern AI datasets and models.

Contribution

It proposes a novel Bias-Guided Generative Network (BGGN) for uncovering intersectional biases, addressing limitations of prior subgroup fairness research.

Findings

BGGN effectively discovers diverse high-bias intersectional attributes.

Generated biased data reveals potential unfairness in AI systems.

Using AI to generate and analyze biased data offers new fairness insights.

Abstract

AI systems have been shown to produce unfair results for certain subgroups of population, highlighting the need to understand bias on certain sensitive attributes. Current research often falls short, primarily focusing on the subgroups characterized by a single sensitive attribute, while neglecting the nature of intersectional fairness of multiple sensitive attributes. This paper focuses on its one fundamental aspect by discovering diverse high-bias subgroups under intersectional sensitive attributes. Specifically, we propose a Bias-Guided Generative Network (BGGN). By treating each bias value as a reward, BGGN efficiently generates high-bias intersectional sensitive attributes. Experiments on real-world text and image datasets demonstrate a diverse and efficient discovery of BGGN. To further evaluate the generated unseen but possible unfair intersectional sensitive attributes, we formulate them as prompts and use modern generative AI to produce new texts and images. The results of frequently generating biased data provides new insights of discovering potential unfairness in popular modern generative AI systems. Warning: This paper contains generative examples that are offensive in nature.