Popping the Social Bubble: Using AI to Increase Transparency in Harmful Content Moderation
摘要
Content moderation in Online Social Networks (OSNs) is an increasingly complex task due to the growth of generated content resulting from the massive increase in the user base of such services. The crucial aspect of the problem has led big companies to hire specialised teams in order to make the service (i.e., the Online Social Network) accessible to potentially anyone; furthermore, in some cases, this task is performed under a nondisclosure agreement due to the very disturbing nature of such contents [1, 2].