Vicarious Offense and Noise Audit of Offensive Speech Classifiers: Unifying Human and Machine Disagreement on What is Offensive

Tharindu Weerasooriya, Sujan Dutta, Tharindu Ranasinghe, Marcos Zampieri, Christopher Homan, Ashiqur KhudaBukhsh


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.
Anthology ID:
2023.emnlp-main.713
Original:
2023.emnlp-main.713v1
Version 2:
2023.emnlp-main.713v2
Volume:
Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing
Month:
December
Year:
2023
Address:
Singapore
Editors:
Houda Bouamor, Juan Pino, Kalika Bali
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
11648–11668
Language:
URL:
https://aclanthology.org/2023.emnlp-main.713
DOI:
10.18653/v1/2023.emnlp-main.713
Bibkey:
Cite (ACL):
Tharindu Weerasooriya, Sujan Dutta, Tharindu Ranasinghe, Marcos Zampieri, Christopher Homan, and Ashiqur KhudaBukhsh. 2023. Vicarious Offense and Noise Audit of Offensive Speech Classifiers: Unifying Human and Machine Disagreement on What is Offensive. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 11648–11668, Singapore. Association for Computational Linguistics.
Cite (Informal):
Vicarious Offense and Noise Audit of Offensive Speech Classifiers: Unifying Human and Machine Disagreement on What is Offensive (Weerasooriya et al., EMNLP 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.emnlp-main.713.pdf
Video:
 https://aclanthology.org/2023.emnlp-main.713.mp4