Reverse-Engineering Twitter's Content Removal
Total Page:16
File Type:pdf, Size:1020Kb
“We Believe in Free Expression...” Reverse-Engineering Twitter’s Content Removal Policies for the Alt-Right The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:38811534 Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA Contents The Problem & The Motivation .............................................................................. 4 Free Speech: Before and After the Internet ......................................................... 5 Speech on Twitter .............................................................................................. 11 Defining the Alt-Right ....................................................................................... 13 The Alt-Right on Social Media ......................................................................... 14 Social Media Reaction to Charlottesville .......................................................... 17 Twitter’s Policies for the Alt-Right ................................................................... 19 Previous Work ................................................................................................... 21 Structure of this Thesis ...................................................................................... 22 Data Collection Methodology ............................................................................... 24 Looking to King and Berger’s Research ........................................................... 24 i. Preliminary Findings ............................................................................... 26 ii. Pivot to Seed Accounts ........................................................................... 28 Overarching Data Collection Pipeline ............................................................... 29 Technological Constraints ................................................................................. 31 Data Collection Pipeline in Detail ..................................................................... 33 Data Collected ................................................................................................... 39 Analysis & Results ................................................................................................ 40 Getting a Sense of the Alt-Right on Twitter ...................................................... 41 Removed versus Active Users ........................................................................... 44 Successful Policy Implementations ................................................................... 47 1. Promotion of the Alt-Right via Propaganda ........................................... 47 2. Representation of the Alt-Right .............................................................. 57 3. Maintaining Free Expression .................................................................. 60 Failure to Meet Policies ..................................................................................... 62 1. Overreach: Awareness of Suspension ..................................................... 62 2. Violation: Failure to Remove Popular Alt-Right Profiles ...................... 64 3. Violation: Failure to Stop Recruitment .................................................. 67 Summary of Findings ........................................................................................ 74 Conclusion & Future Research .............................................................................. 75 Speculation of Political Motivation ................................................................... 78 Twitter Removal of Antifa ................................................................................ 79 Further Research ................................................................................................ 82 In Conclusion: The Future of Speech on Twitter .............................................. 83 Bibliography .......................................................................................................... 87 ! 3 1 The Problem & The Motivation On October 11, 2017, at approximately seven pm Eastern Standard Time, President Donald Trump’s Twitter account was deactivated. People searching for @realDonaldTrump were greeted with a banner: “Sorry, this page doesn’t exist!”2 A mere eleven minutes later, the Commander-in-Chief’s page was reactivated.3 It might have seemed like a comical fluke – and, for many, a relief – but this removal was the work of a disgruntled Twitter employee on his last day of work.4 Given that many argue that Twitter offers direct access to the President’s voice, it is notable that a single employee was able to remove his account.5 This event raised a host of questions about Twitter’s content removal process, many of which !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 2 Maggie Astor, “Rogue Twitter Employee Briefly Shuts Down Trump’s Account,” New York Times, November 2, 2017, https://www.nytimes.com/2017/11/02/us/politics/trump-twitter-deleted.html. 3 Ibid. 4 Ibid. 5 Jameel Jaffer, “Government Secrecy in the Age of Information Overload,” Salant Lecture, Shorenstein Center, Cambridge, MA, October 17, 2017. ! 4 went largely unanswered.6 The problem is not Twitter’s alone; for most social media companies, despite published content policies, there are no statistics or publicly available reasoning regarding removal. Given that global society is increasingly reliant on these platforms to communicate and connect, this lack of transparency in content regulation is unsettling. How did social media platforms like Twitter evolve to a point where unchecked content removal is not unusual, and why is it important that these companies clearly define a sense of criteria and precedence for the content they remove? Free Speech: Before and After the Internet ! Free speech is deeply rooted in America’s identity, as evidenced by its enshrinement in the First Amendment of the Constitution.7 However, given that companies like Google, Facebook, and Twitter are private corporations, they are under no obligation to follow this model of free speech; they have the right to determine what content should stay or be removed on their platforms. To better understand the effects of these companies on public discourse, this thesis will begin by discussing the notion of free speech beginning two thousand years ago. The concept of free speech is intertwined with the concept of privacy, which was first pioneered by Greek philosopher Aristotle. Aristotle envisioned privacy as a separation of the public life of the agora from the private life of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 6 Astor, “Rogue Twitter Employee.” 7 The First Amendment: “Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof; or abridging the freedom of speech, or of the press; or the right of the people peaceably to assemble, and to petition the government for a redress of grievances.” ! 5 home.8,9 According to Aristotle, in the agora, the public space, it is assumed that whatever one says can be heard by others, whereas in the home, it is assumed that one’s words are private. However, the invention of new technology altered this distinction between public and private life, intertwining the notions of privacy and free speech. In 1888, the invention of the Kodak camera sparked a great deal of panic. While previously, photos required the subject to sit still to be captured, the Kodak camera allowed instant photography. Photographers began taking pictures of people in public, going about their every day lives, without their permission. 10 Soon, concern grew about the potential risk to the safety of personal lives and reputations. This prompted Justice Louis Brandeis to co-author the famous essay, “The Right to Privacy,” which argued that, unlike European law, American constitutional law contains no protection of speech (or photography) that results in an “offense against honor.”11 In the decision for the Supreme Court case Whitney vs. California (1927), Brandeis clarified his understanding of free speech by referencing Thomas Jefferson and the revolutionaries of 1776: [Those who won our independence] knew that order cannot be secured merely through fear of punishment for its infraction; that it is hazardous to discourage thought, hope, and imagination; that fear breeds repression; that repression breeds hate; that hate !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 8 Judith DeCew, “Privacy,” Stanford Encyclopedia of Philosophy, Stanford University, 14 May 2002, https://plato.stanford.edu/entries/privacy/. 9 Agora: from ancient Greece, a public space used for assemblies and markets. 10 Steve Lohr, “With 'Brandeis' Project, Darpa Seeks to Advance Privacy Technology,” The New York Times, September 14, 2015, https://bits.blogs.nytimes.com/2015/09/14/with-brandeis-project-darpa-seeks-to- advance-privacy-technology/. 11 Samuel Warren and Louis Brandeis, “The Right to Privacy,” Harvard Law Review, vol. 4, no. 5, (1890): pp. 193–220. ! 6 menaces stable government; that the path of safety lies in the opportunity to discuss freely supposed grievances and proposed remedies; and that the fitting remedy for evil counsels is good ones.12 Jefferson’s understanding