Vicarious Offense and Noise Audit of Offensive Speech Classifiers: Unifying Human and Machine Disagreement on What is Offensive
arXiv:2301.12534 · doi:10.18653/v1/2023.emnlp-main.713
Abstract
Offensive speech detection is a key component of content moderation. However, what is offensive can be highly subjective. This paper investigates how machine and human moderators disagree on what is offensive when it comes to real-world social web political discourse. We show that (1) there is extensive disagreement among the moderators (humans and machines); and (2) human and large-language-model classifiers are unable to predict how other human raters will respond, based on their political leanings. For (1), we conduct a noise audit at an unprecedented scale that combines both machine and human responses. For (2), we introduce a first-of-its-kind dataset of vicarious offense. Our noise audit reveals that moderation outcomes vary wildly across different machine moderators. Our experiments with human moderators suggest that political leanings combined with sensitive issues affect both first-person and vicarious offense. The dataset is available through https://github.com/Homan-Lab/voiced.
Accepted at EMNLP 2023
References in corpus (8)
- Training language models to follow instructions with human feedback
- Llama 2: Open Foundation and Fine-Tuned Chat Models
- What is Gab? A Bastion of Free Speech or an Alt-Right Echo Chamber?
- #Election2020: The First Public Twitter Dataset on the 2020 US Presidential Election
- Hate Towards the Political Opponent: A Twitter Corpus Study of the 2020 US Elections on the Basis of Offensive Speech and Stance Detection
- Learning Soft Labels via Meta Learning
- Same Words, Different Meanings: Semantic Polarization in Broadcast Media Language Forecasts Polarization on Social Media Discourse
- Are Chess Discussions Racist? An Adversarial Hate Speech Data Set