The Disparate Effects of Partial Information in Bayesian Strategic Learning

2025-05-31

AI Summary

This paper investigates how information opacity in scoring rules—where agents receive only noisy signals about the underlying decision rule—affects group fairness. We develop a Bayesian updating and strategic response game model to characterize asymmetric behavioral responses across groups under heterogeneous feature-modification costs and prior beliefs. Our key contributions are threefold: (1) transparency and fairness exhibit a non-monotonic relationship, with moderate transparency often yielding optimal fairness; (2) Bayesian-rational agents mitigate explosive outcome disparities, whereas naive agents exacerbate harm to low-cost groups; and (3) we derive, for the first time, an analytical functional form characterizing how group disparity varies with transparency, and establish a provably bounded upper bound on fairness gaps under heterogeneous priors. These results provide theoretical foundations and quantitative guidance for transparency design in algorithmic governance.

Technology Category

• Game Theory and Economic Paradigms: Fair Division
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Philosophy and Ethics of AI: Bias, Fairness & Equity

Application Category

• Security and Privacy: Data transparency and provenance
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments
• Responsible Web: Algorithmic accountability and transparency on the web

Abstract

We study how partial information about scoring rules affects fairness in strategic learning settings. In strategic learning, a learner deploys a scoring rule, and agents respond strategically by modifying their features -- at some cost -- to improve their outcomes. However, in our work, agents do not observe the scoring rule directly; instead, they receive a noisy signal of said rule. We consider two different agent models: (i) naive agents, who take the noisy signal at face value, and (ii) Bayesian agents, who update a prior belief based on the signal. Our goal is to understand how disparities in outcomes arise between groups that differ in their costs of feature modification, and how these disparities vary with the level of transparency of the learner's rule. For naive agents, we show that utility disparities can grow unboundedly with noise, and that the group with lower costs can, perhaps counter-intuitively, be disproportionately harmed under limited transparency. In contrast, for Bayesian agents, disparities remain bounded. We provide a full characterization of disparities across groups as a function of the level of transparency and show that they can vary non-monotonically with noise; in particular, disparities are often minimized at intermediate levels of transparency. Finally, we extend our analysis to settings where groups differ not only in cost, but also in prior beliefs, and study how this asymmetry influences fairness.

Problem

• Examining fairness disparities in strategic learning with partial scoring rule information
• Analyzing naive vs Bayesian agents' response to noisy scoring rule signals
• Investigating how transparency levels affect outcome disparities across cost-differing groups

Innovation

• Studying partial noisy signals in scoring rules
• Comparing naive and Bayesian agent behaviors
• Analyzing disparities with varying transparency levels