Loading...
Paper Type
Complete
Description
Content moderation is a common intervention strategy for reviewing user-generated content on social media platforms. Engaging users in content moderation is promising for making ethical and fair moderation decisions. A few studies that have considered user engagement in content moderation have primarily focused on classifying user-generated comments, rather than leveraging the information of user engagement to make a moderation decision on user-generated posts. Moreover, how to extract information from user engagement to enhance content moderation remains unclear. To address the above-mentioned limitations, this study proposes a framework for user engagement-enhanced moderation of user-generated posts. Specifically, it incorporates the credibility and stance of user-generated content into graph learning. Our empirical evaluation shows that the models based on our proposed framework outperform the state-of-the-art deep learning models in making moderation decisions for user-generated posts. The findings of this study have implications for augmenting the moderation of social media content and for improving the safety and success of online communities.
Paper Number
1175
Recommended Citation
Wang, Kanlun; Fu, Zhe; Zhou, Lina; and Zhang, Dongsong, "How Does User Engagement Support Content Moderation? A Deep Learning-based Comparative Study" (2023). AMCIS 2023 Proceedings. 3.
https://aisel.aisnet.org/amcis2023/sig_aiaa/sig_aiaa/3
How Does User Engagement Support Content Moderation? A Deep Learning-based Comparative Study
Content moderation is a common intervention strategy for reviewing user-generated content on social media platforms. Engaging users in content moderation is promising for making ethical and fair moderation decisions. A few studies that have considered user engagement in content moderation have primarily focused on classifying user-generated comments, rather than leveraging the information of user engagement to make a moderation decision on user-generated posts. Moreover, how to extract information from user engagement to enhance content moderation remains unclear. To address the above-mentioned limitations, this study proposes a framework for user engagement-enhanced moderation of user-generated posts. Specifically, it incorporates the credibility and stance of user-generated content into graph learning. Our empirical evaluation shows that the models based on our proposed framework outperform the state-of-the-art deep learning models in making moderation decisions for user-generated posts. The findings of this study have implications for augmenting the moderation of social media content and for improving the safety and success of online communities.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
SIG AIAA