Fairness for the People, by the People: Minority Collective Action
This addresses fairness issues for minority groups in ML applications, offering a user-driven alternative to firm-side methods, though it is incremental as it builds on existing algorithmic collective action frameworks.
The paper tackles the problem of unfair treatment of minority groups in machine learning models by proposing that a coordinated minority group can strategically relabel its own data to enhance fairness without altering the firm's training process, and validation on real-world datasets shows that a subgroup can substantially reduce unfairness with a small impact on overall prediction error.
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.