Healthy deliberation is a crucial means to enhance social cohesion and mutual understanding in society, by learning from other viewpoints and collaboratively developing new ones.
Through digital connectivity, people with diverse backgrounds can potentially find each other on comment sections in news platforms and engage in constructive and inclusive debate. Yet, this potential of inclusive deliberation has not been realized to date. While news platforms manage to block the bulk of uncivil posts and highlight the most constructive ones, by means of a fusion between machine learning and intensive human moderation, the current practice lacks a transparent moderation for the more complex discussion topics and does not give equal attention to minority viewpoints due to the sheer volume of other messages. To address these shortcomings, and move forward to the potential of inclusive deliberation, we propose to develop a tool that summarizes any set of comments posted in the context of news reports, by highlighting the topic and degree of constructiveness of single posts and visualizing the output in a manner that invites visitors into discussion and informs them with a heterogeneous overview of the topics and viewpoints that have been shared. Our project will provide novel insights into computer-assisted forum moderation by bringing together an innovative team of researchers, moderators and media-stakeholders and studying the postulations of particular discussion topics. We will contribute to insights on heterogeneous discussion visualization by adapting the visualizations to the structure of a discussion at hand and comparing closed-domain and open-domain approaches.