Towards Effective Discrimination Testing for Generative AI
This addresses the challenge of regulating discriminatory behavior in generative AI systems, which is crucial for ensuring fairness in deployments, though it is incremental as it builds on existing legal and technical literature.
The paper tackles the problem of ineffective discrimination testing for generative AI models, showing through case studies that misalignment between fairness testing techniques and regulatory goals can lead to discriminatory outcomes in real-world deployments.
Generative AI (GenAI) models present new challenges in regulating against discriminatory behavior. In this paper, we argue that GenAI fairness research still has not met these challenges; instead, a significant gap remains between existing bias assessment methods and regulatory goals. This leads to ineffective regulation that can allow deployment of reportedly fair, yet actually discriminatory, GenAI systems. Towards remedying this problem, we connect the legal and technical literature around GenAI bias evaluation and identify areas of misalignment. Through four case studies, we demonstrate how this misalignment between fairness testing techniques and regulatory goals can result in discriminatory outcomes in real-world deployments, especially in adaptive or complex environments. We offer practical recommendations for improving discrimination testing to better align with regulatory goals and enhance the reliability of fairness assessments in future deployments.