Proceedings of the Third Workshop on Abusive Language Online

The last few years have seen a surge in attention to various forms of abuse such as cyberbullying, hate speech, and scapegoating occurring on online platforms. At the same time, there has been a rise in interest in using Natural Language Processing (NLP) to address these issues at scaale. However, in order to develop robust, long-term solutions for this problem, we require perspectives from diverse disciplines ranging from psychology, law, gender studies, communications, and critical race theory. Our goal with the Abusive Language Workshop is to provide a platform to facilitate the interdisciplinary conversations and collaborations necessary to thoughtfully address the issue of abuse at scale.Each year, we choose a theme for our workshop that guides the talks and panel discussions at the workshop. In previous years we focused on the policy aspect of online abuse and the stories and experiences of those who have received large amounts of online abuse. The themes do not limit the original research presented at the workshop, rather it helps frame the research presented through the lens of its potential to address the concerns of the theme. For this year, we have chosen to focus on human content rating, the practice of annotating and moderating data-an aspect which is often unspoken, assumed, and often forms the basis of the research conducted.