Towards a Normative Ethics of Online Moderation

• In Event: Critical Data Studies: Ethics and Human Contexts of Data Science III

Abstract

The tension between inclusion and exclusion is an essential aporia of democracy. Liberal democracy values inclusion and diversity, but it tempers this with a condition of mutual respect (Gutmann and Thomspon, 1996). More-radical or critical approaches push inclusion further (Young, 2000), or reject the constraint of mutual respect entirely (Fish, 1999). These ethico-political questions also apply to digital publics and online content moderation, particularly automated moderation. This paper aims to outline the key ethical questions both practitioners and scholars should consider in designing moderation policies and systems. Some questions include: what categories of online behavior should be used to determine whether an individual should be sanctioned or banned from a group? Metadata? If so, what kinds? Associations or behavior in other groups? Additionally, what weight should we give to what actions – aggressive speech, harassment, hate speech, sexist/racist/ableist language? Can any users be banned or silenced a priori: e.g. for affiliation with hate groups? Should bots have the same rights/responsibilities as human users? How long should a ban or mark for abusive behavior stay with a user – can their be hope for rehabilitation or “paying one’s dues”? What about users that use obfuscation to mask behavior and identity in community, and across platforms? What should the balance be between automated and human moderation? Should there be “due process”? Engagement with these questions, we hope, will move us towards a normative ethical framework for online moderation, and also contribute to contemporary theory about digital democracy more broadly.

Authors

• Andrew Arthur Fitzgerald, Stanford University

• Anna Gibson