Research

Should the Powerful Be Moderated? Examining Social Media Users’ Perspectives on Moderating Politicians’ Inappropriate Speech

Platforms' capitalistic functions
and free speech concerns
influence user attitudes

Alyvia Walters and Shagun Jhaver (2026), “Should the Powerful Be Moderated? Examining Social Media Users’ Perspectives on Moderating Politicians’ Inappropriate Speech,” Accepted in Proc. ACM Hum.-Comput. Interact., CSCW.


Abstract

Prior HCI research has investigated social media users’ perspectives on the efficacy and fairness of content moderation tools, as well as their concerns around the ways that powerful individuals may not be subject to the same “rules” as other users. However, we argue that there is a sub-group of powerful accounts that deserve specialized attention: those of politicians. Through a focus group study, we outline user perspectives on how—or if—politicians’ inappropriate online speech should be moderated, and what mental models are influencing these perspectives. We find that participants’ concerns about platforms’ capitalistic functions, the changed relationship between politics and media, the outsized influence of politicians’ speech, and First Amendment considerations all play heavily into their conceptions of the risks and benefits of moderating these powerful accounts. We also examine how participants perceive the benefits and shortcomings of different platform-enacted and user-directed moderation solutions to counteract inappropriate politician speech. Building upon our findings, we offer suggestions on how scholars of law and platform governance can help build design and policy solutions that better address this issue. We also discuss how HCI research can contribute to advancing the debates around it.

BibTeX citation

@article{walters-2026-politicians,
    author = {Walters, Alyvia and Jhaver, Shagun}, 
    title = {Should the Powerful Be Moderated? Examining Social Media Users’ Perspectives on Moderating Politicians’ Inappropriate Speech}, 
    year = {2026}, 
    journal = {Proc. ACM Hum.-Comput. Interact.}, 
    }