Primary Submission Category: Causal Fairness, and Bias/Discrimination
The IEEE P3591 Standard for Fair Decision Making Through Causal Analysis
Authors: Christopher Lam,
Presenting Author: Christopher Lam*
Causal inference is increasingly used to study algorithmic bias, fairness, and discrimination in high-stakes decision systems, yet the field lacks a shared, operational standard for representing these concepts in causal terms. This gap is often filled by statistical fairness metrics that are incompatible with one another and poorly aligned with legal and regulatory requirements.
IEEE P3591 is an emerging international standard that defines fair decision making as a property of causal structure rather than statistical outcomes. The standard introduces a causal ontology for protected attributes, mediators, decisions, and outcomes, and specifies how permissible and impermissible causal pathways can be identified and documented. IEEE P3591 is designed to bridge causal inference research and real-world AI deployment while preserving methodological flexibility, providing a common causal language for policymakers, lawyers, economists, and data scientists.