Abusive Comments in Online Media and How to Fight Them: State of the Domain and a Call to Action
Niemann Marco, Welsing Jens, Riehle Dennis M, Brunk Jens, Assenmacher Dennis, Becker Jörg
Abstract
While abusive language in online contexts is a long-known problem, algorithmic detection and moderation support are only recently experiencing rising interest. This survey provides a structured overview of the latest academic publications in the domain. Assessed concepts include the used datasets, their language, annotation origins and quality, as well as applied machine learning approaches. It is rounded off by an assessment of meta aspects such as author collaborations and networks as well as extant funding opportunities. Despite all progress, the domain still has the potential to improve on many aspects: (international) collaboration, diversifying and increasing available datasets, careful annotations, and transparency. Furthermore, abusive language detection is a topic of high societal relevance and requires increased funding from public authorities.
Keywords
Abusive language; Comment moderation; Machine learning; Review