온라인 토론의 부정에 대한 피드백 메시지에 대한 사용자의 반응 이해

Abstract

A common problem of online discussion is harassment where harassment may be political, sexual, or racial. Existing moderation methods focus on detecting harassment contents and taking action when appropriate. In spite of the availability of different detection and moderation techniques, online harassment is not an issue which can be reduced easily. Mostly missing in the existing research is designing to encourage users in the behavior which align with the normative behavior of the community, and reduce uncivil content. In this work, I explore the potential of promoting civil commenting, by providing feedback messages to the user regarding the existence of incivility in user generated content

and investigate user’s reaction and future posting behavior after the feedback message. How do users react if providing the machine learning output of having potential negative content in their post? How do different components of the feedback message affect users' reactions? Is there any positive effect in the future commenting behavior of sending these feedback messages? Analyzing the answers of the above questions can offer new opportunities to significantly reduce harassment content online, by proactively notifying users about the potential bad contents in their post. The result from this work, shows promising potential, which needs to investigate more in this aspect.