META’S 3PFC SPEECH GOVERNANCE: AN INQUIRY INTO THE FACT-CHECKING CONTENT MODERATION INFRASTRUCTURE

Abstract

Social media play a pivotal role in managing the environments where political debate and public deliberation occur, and therefore must contend with the targets set by official bodies and policymakers to ward off harmful speech and mis/disinformation in their platforms. However, little is known about the criteria applied by social platforms’ content moderation infrastructure to address problematic information. This paper examines the speech governance parameters applied by Meta’s 3PFC to moderate problematic information. Leveraging digital methods and approaches developed in journalism studies, it implements a series of manual and computational techniques to curate a comprehensive dataset combining fact-checking content commissioned by Meta’s 3PFC program—available through the Facebook URLs Dataset (Meta, n.d.)—and fact-checks produced independently from Meta’s program. We probe this database to identify the criteria applied by Meta’s speech governance through the 3PFC program across five countries: Argentina, Philippines, Portugal, United Kingdom, and South Africa, with content spanning three languages (English, Spanish, and Portuguese). Considering the looming end of Meta’s 3PFC system worldwide and the recently declared laissez-faire commitments of US-based tech companies to public speech (Silverman, 2025), this paper addresses the normative standards consolidated by Meta’s content moderation infrastructure through outsourced fact-checking work. The findings contribute to developing platform governance policies in a context where countries like Australia, Brazil, India, and the European Union are designing or implementing regulatory frameworks to hold social platforms accountable (Anastácio, 2024; Liu, 2024; Ó Fathaigh et al., 2021).