Facts and Where to Find Them: Empirical Research on Internet Platforms and Content Moderation

Abstract

We live in an era of increasing worry that internet platforms like Facebook or Twitter, which mediate our online speech, are also fomenting hatred, spreading misinformation, and distorting political outcomes. The 2016 US presidential election, in particular, unleashed a torrent of concern about platform-borne harms. Policymakers around the world have called for laws requiring platforms to do more to combat illegal and even merely "harmful" content.From the perspective of platforms themselves, these proposals have a lot in common. Regardless of their substantive mandatesto address content that is misleading, hateful, or violent, for examplethey all require similar operational processes to comply. Platforms already have these processes in place to enforce current laws and their discretionary Community Guidelines. Any new efforts to regulate content online will likely build on existing systems, personnel, and toolsand inherit both their strengths and their weaknesses. That makes it important to understand those systems.Reliable information about platforms' content-removal systems was, for many years, hard to come by; but data and disclosures are steadily emerging as researchers focus on the topic and platforms ramp up their transparency efforts. This chapter reviews the current and likely future sources of information.Some content takedowns are required by law, while others are performed voluntarily. Legal takedowns are shaped by intermediary liability laws, which tell platforms what responsibility they have for unlawful content posted by

Clandestino or Rifugiato?Anti-immigration Facebook Ad Targeting in Italy✱

“…In fact, as private companies, they are free to enforce any internal policy they choose [53]. Interviews with insiders and employees of these companies [40,42] show that these policies are determined via "internal debates, appeals by practitioners, and outside pressure" and are applied in opaque ways. Unfortunately, the Library does not reveal the advertisements that have been rejected by Facebook.…”

Moderating manipulation: Demystifying extremist tactics for gaming the (regulatory) system

“…However, research has shown that a consequence of stringent filtering systems and censorship is that those who have been deplatformed from mainstream sites may migrate to private chats and encrypted sites like Telegram (Bloom et al, 2019), which, although often smaller and offering extremists a reduced audience, are perceived by users as more secretive through their encoding, making it far more difficult for public monitoring of misuse and emerging threats. Concern has also arisen with respect to the possibility of subjective, disparate and potentially biased enforcement of content and creators by different platforms resulting from the absence of internationally agreed upon definitions of what constitutes terrorism and violent extremism (Keller & Leerssen, 2020). Moreover, transparency regarding the tools and approaches used by social media companies for content moderation is highly limited and the absence of details regarding their implementation makes evaluating their effectiveness difficult (United Nations, 2021).…”

Content Moderation As a Political Issue: The Twitter Discourse Around Trump's Ban

“…For the most part, moderators review content only after it has been flagged and brought to their attention, either algorithmically or by users themselves (Klonick, 2018). Yet, mounting pressure from governments and public groups in recent years has nudged platforms to adopt a more proactive approach to the detection and removal of content that violates their own private rules or public laws (Keller et al, 2020;Gillespie, 2020).…”