There is a toolbox of content moderation options available to online platforms such as labeling, algorithmic sorting, and removal. A content label is a visual and/or textual attachment to a piece of user-generated content intended to contextualize that content for the viewer. Examples of content labels are fact-checks or additional information. At their essence, content labels are simply information about information. If a social media platform decides to label a piece of content, how does the current body of social science inform the labeling practice? Academic research into content labeling is nascent, but growing quickly; researchers have already made strides toward understanding labeling best practices to deal with issues such as misinformation, conspiracy theories, and misleading content that may affect everything from voting to personal health. We set aside normative or ethical questions of labeling practice, and instead focus on surfacing the literature that can inform and contextualize labeling effects and consequences. This review of a kind of “emerging science” summarizes the labeling literature to date, highlights gaps for future research, and discusses important considerations for social media platforms. Specifically, this paper discusses the particulars of content labels, their presentation, and the effects of various label formats and characteristics. The current literature can help guide the usage and improvement of content labels on social media platforms and inform public debate and policy over platform moderation.

Keywords: social media, content moderation, misinformation, First Amendment, free expression, technology platforms


Read the full paper here.