Abstract

How, if at all, should governments be involved in prescribing content moderation policies for extreme speech on social media platforms? There are a range of regulatory models that may be adopted, on a spectrum from laissez-faire self-regulation to punitive state control. We propose a normative taxonomy of such schemes, that is, a classification based on how the different regulatory models bear upon the normative legitimacy of state (in)action in this area. We defend a certain kind of hybrid regulatory model (i.e., government working with platforms on regulatory policy), which is legislative, rather than informal or advisory, so as to ensure democratic input, transparency, accountability and redress (due process), but which imposes procedural requirements upon companies’ moderation policies and practices, rather than punitively prescribing moderation outcomes in specific cases.