From Individual Experience to Collective Evidence: A Reporting-Based Framework for Identifying Systemic Harms

2025-02-12

AI Summary

This study aims to contextualize individual adverse system interaction reports—such as loan denials or vaccine adverse events—within systemic behavioral patterns to identify structurally vulnerable subpopulations disproportionately affected by systemic harm. Method: We formulate systematic harm detection as a sequential hypothesis testing problem with multiple testing correction, introducing the “reporting behavior inferability” condition to ensure statistical reliability of true harm-rate differences across subpopulations. Our approach integrates Sequential Probability Ratio Testing (SPRT), Benjamini–Hochberg false discovery rate (FDR) control, and adaptive subgroup composition search, grounded in real-world event databases. Contribution/Results: Evaluated on mortgage approval and vaccine adverse event data, our method accurately recovers known vulnerable subgroups using only 15–30% of the original data volume. It demonstrates high efficiency, robustness to reporting heterogeneity, and full interpretability—establishing the first statistically rigorous framework for detecting systemic inequities from sparse, noisy, individual-level adverse reports.

Technology Category

• Multiagent Systems: Adversarial Agents
• Reasoning under Uncertainty: Causality
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Large-scale security measurements

Abstract

When an individual reports a negative interaction with some system, how can their personal experience be contextualized within broader patterns of system behavior? We study the incident database problem, where individual reports of adverse events arrive sequentially, and are aggregated over time. In this work, our goal is to identify whether there are subgroups--defined by any combination of relevant features--that are disproportionately likely to experience harmful interactions with the system. We formalize this problem as a sequential hypothesis test, and identify conditions on reporting behavior that are sufficient for making inferences about disparities in true rates of harm across subgroups. We show that algorithms for sequential hypothesis tests can be applied to this problem with a standard multiple testing correction. We then demonstrate our method on real-world datasets, including mortgage decisions and vaccine side effects; on each, our method (re-)identifies subgroups known to experience disproportionate harm using only a fraction of the data that was initially used to discover them.

Problem

• Identify subgroups experiencing systemic harms.
• Contextualize individual reports into collective evidence.
• Apply sequential hypothesis tests to detect disparities.

Innovation

• Sequential hypothesis testing applied
• Identifies disparities in harm rates
• Uses multiple testing correction algorithms