How Reliable are Fairness Audits with Unreliable Data?

Title:How Reliable are Fairness Audits with Unreliable Data?

Abstract:Fairness audits are a key component of responsible machine-learning deployment. Yet, the reliability of audit recommendations under incomplete protected-label access is still poorly understood. In this work, we focused on protected-label missingness in fairness mitigation audits. We introduced a seed-calibrated stress test to separate missingness effects from seed-to-seed movement that is already present under complete labels. Across ACS/Folktables tasks, we found that positive-availability missingness usually does not move selected mitigation methods beyond the complete-label seed floor. The no-label endpoint behaves differently, exposing ERM-equivalent candidates and deterministic tie-breaking rather than a broad missingness effect. We also found that threshold optimization can turn single-axis fairness gains into above-null intersectional harm, a sharper failure pattern that appears to remain visible under random-forest validation. Overall, our results highlight that protected-label missingness should be reported with seed-null calibration, candidate-set context, and intersectional consequences before it is treated as evidence of audit fragility.

Machine Learning (cs.LG); Machine Learning (stat.ML)

arXiv:2506.23033 [cs.LG]

(or arXiv:2506.23033v2 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2506.23033 Focus to learn more arXiv-issued DOI via DataCite

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.