Fairness for the People, by the People: Minority Collective Action

TL;DR

This paper introduces a novel framework called Algorithmic Collective Action, enabling minority groups to improve fairness in machine learning models by strategically relabeling their data, without changing the model training process.

Contribution

It proposes three practical, model-agnostic methods for minority groups to induce fairness through data relabeling, validated on real-world datasets.

Findings

Minority groups can significantly reduce unfairness

Small impact on overall prediction error

Framework is model-agnostic and practical

Abstract

Machine learning models often preserve biases present in training data, leading to unfair treatment of certain minority groups. Despite an array of existing firm-side bias mitigation techniques, they typically incur utility costs and require organizational buy-in. Recognizing that many models rely on user-contributed data, end-users can induce fairness through the framework of Algorithmic Collective Action, where a coordinated minority group strategically relabels its own data to enhance fairness, without altering the firm's training process. We propose three practical, model-agnostic methods to approximate ideal relabeling and validate them on real-world datasets. Our findings show that a subgroup of the minority can substantially reduce unfairness with a small impact on the overall prediction error.