analiza aiqbench
This research clarifies how fairness audit metrics behave when protected demographic labels are missing or incomplete. The authors demonstrate that without proper baselines and candidate set context, reported rates can be misleadingly interpreted as evidence of model bias rather than artifacts of data limitations. For practitioners deploying ML systems, the key takeaway is to always report these specific contextual details alongside any published fairness statistics before drawing conclusions about a deployed model's safety or equity.
What Must a Fairness Audit Report When Demographic Data Is Incomplete?
arXiv:2506.23033v5 Announce Type: replace Abstract: Fairness audits are a key component of responsible machine-learning deployment. Yet what such an audit must disclose, when the protected labels it depends on are incomplete, remains unsettled. In this work, we focused on the rates a fairness audit publishes and on what an oversight reader needs beside them. We paired every published rate with a matched baseline drawn from the same audit, one hiding protected labels and one varying only the run seed. Across ACS/Folktables tasks, missingness settings that kept some protected labels moved the selected mitigation less than an ordinary rerun did. At zero protected-label access, candidates collapsed to empirical risk minimization, so the apparent exception there reflected the candidate set's composition. Equalized-odds threshold optimization most often regressed an intersectional subgroup, but that rate fell back to its baseline once we kept only the configurations an auditor would accept. Any accuracy it lost fell on the population as heavily as on the worst-off cell, so the mechanism is levelling down. The one effect that survived was a change in which cell is worst-off. Overall, our results highlight that a published audit rate should be reported with the baseline needed to interpret it, the candidate set it came from, and its intersectional effects, before it is treated as evidence about a deployed model.
Abstract:Fairness audits are a key component of responsible machine-learning deployment. Yet what such an audit must disclose, when the protected labels it depends on are incomplete, remains unsettled. In this work, we focused on the rates a fairness audit publishes and on what an oversight reader needs beside them. We paired every published rate with a matched baseline drawn from the same audit, one hiding protected labels and one varying only the run seed. Across ACS/Folktables tasks, missingness settings that kept some protected labels moved the selected mitigation less than an ordinary rerun did. At zero protected-label access, candidates collapsed to empirical risk minimization, so the apparent exception there reflected the candidate set's composition. Equalized-odds threshold optimization most often regressed an intersectional subgroup, but that rate fell back to its baseline once we kept only the configurations an auditor would accept. Any accuracy it lost fell on the population as heavily as on the worst-off cell, so the mechanism is levelling down. The one effect that survived was a change in which cell is worst-off. Overall, our results highlight that a published audit rate should be reported with the baseline needed to interpret it, the candidate set it came from, and its intersectional effects, before it is treated as evidence about a deployed model.
Machine Learning (cs.LG); Machine Learning (stat.ML)
arXiv:2506.23033 [cs.LG]
(or arXiv:2506.23033v5 [cs.LG] for this version)
Text integral sub licența deschisă a publicației · 228 cuvinte
Subiecte conexe
AI in healthcare: applications and best practices
Databricks · 24 iul. 2026
Why I Left Google DeepMind
LessWrong · 15 iul. 2026
AI Transparency: Governance, Explainability, and Data Practices
Databricks · 20 iul. 2026
A Red Line and Oversight Framework for Government AI Contracts
AI Alignment Forum · 18 iul. 2026