Whose Preferences? Differences in Fairness Preferences and Their Impact on the Fairness of AI Utilizing Human Feedback
This addresses fairness in AI content moderation by revealing how demographic differences in human feedback affect system outcomes, though it is incremental in highlighting these variations.
The study examined fairness preferences in content moderation using human feedback, finding significant gaps based on annotator demographics like race, age, and political stance, and showed that an ensemble classifier balancing different demographic annotations performed better across demographic intersections.
There is a growing body of work on learning from human feedback to align various aspects of machine learning systems with human values and preferences. We consider the setting of fairness in content moderation, in which human feedback is used to determine how two comments -- referencing different sensitive attribute groups -- should be treated in comparison to one another. With a novel dataset collected from Prolific and MTurk, we find significant gaps in fairness preferences depending on the race, age, political stance, educational level, and LGBTQ+ identity of annotators. We also demonstrate that demographics mentioned in text have a strong influence on how users perceive individual fairness in moderation. Further, we find that differences also exist in downstream classifiers trained to predict human preferences. Finally, we observe that an ensemble, giving equal weight to classifiers trained on annotations from different demographics, performs better for different demographic intersections; compared to a single classifier that gives equal weight to each annotation.