Is Social Media a Threat to Democracy? ISSUES: Focusing on Six Key Risks Issue Four: Political capture of platforms cross-border and bot-enabled activity). President Duterte of the Philippines has successfully used social media to control the narrative on the drug war (which has resulted in over 7,000 deaths since mid-201629) and to silence any questions on his tactics.30 Prime Minister Modi of India offers a direct-to-your-phone tweet service as part of the Digital India initiative, and he has used Twitter to make significant policy announcements—such as the recent statement on demonetization, which bypassed the conventional route of first presenting a major new policy to Parliament. The issue is worsened through automation. Automated bots can create the false impression of momentum at scale: highly automated accounts, defined as accounts that tweeted 450 or more times with a related hashtag and user mention, generated close to 18% of all Twitter traffic about the 2016 U.S. presidential election. Researchers at USC recently estimated that “between 9% and 15% of active Twitter accounts are bots”—which could mean that there are up to 50 million bot accounts.31 There are also implications for national sovereignty, as the uncertain geographic origin of bots facilitates their use by foreign governments or groups seeking to intervene in another country’s political process. In some ways this undermines a crucial premise of democracy where processes are set up to give voice to the people. Malevolent bots allow certain voices to be amplified disproportionately to manipulate public sentiment. Moreover, few regulations or protocols protect users against trolling and online abuse, and social media companies are not always quick to—or equipped to— respond globally. The Myanmar ICT for Development Organization (MIDO) found that only 10% of the posts it reported as hate speech on Facebook were ultimately removed.32 Issue Five: Manipulation, micro-targeting, and behavior change Unprecedented personal data captured by social media platforms enables sophisticated psychographic or behavioral profiling and micro-targeting. Social media companies today have troves of data that any government would like to access, and, in some cases, successfully have33 One use case for this micro-targeting is “dark advertising,” in which political advertisements are shown to Facebook users without being subject to traditional regulatory guidelines. As Phil Howard argued in a Washington Post opinion piece, “while we are all talking about ‘fake news,’ we should also be talking about the algorithms and fake accounts that push bad information around.” Technology firms, with large amounts of personal data, support the experiments that can nudge behavior. Michael Bueza, IN NUMBERS: The Philippines’ ‘war on drugs’, Rappler, Updated April 23, 2017, https://www.rappler.com/newsbreak/ iq/145814-numbers-statistics-philippines-war-drugs Rappler is an Omidyar Network investee. 30 Interview with Maria Ressa, Rappler (Philippines), 2017. See also Rappler’s series on the weaponization of speech in the Philippines: http://www.rappler.com/nation/148007-propaganda-warweaponizing-internet. Rappler is an Omidyar Network investee. 31 The researchers used more than one thousand features to identify bot accounts on Twitter, in categories including friends, tweet content and sentiment, and time between tweets. See Social 29 © Copyright The Omidyar Group 2017. All rights reserved. Big data analysis, combined with computational psychology and behavioral and demographic analysis, has allowed firms (not just Google and Facebook, but reportedly also Cambridge Analytica, though the claim of the influence of the latter may be overstated) to “decode” users’ personalities and target them with tailored messaging to drive behavioral shifts. Since the fundamental business model of these platforms is based on monetizing user data and attention, we might expect this to continue, if not intensify. The absence of appropriate regulation that allows widespread surveillance and collection of user data has extended into democratic election campaigns. The continuous surveillance of users through feedback bots distort the 2016 U.S. Presidential election, online discussion by Alessandro Bessi and Emilio Ferrara. First Monday, Volume 21, Number 11 -7 November 2016, http://firstmonday.org/ojs/index.php/ fm/article/view/7090/5653 32 See https://www.buzzfeed.com/meghara/how-fake-news-andonline-hate-are-making-life-hell-for? 33 Glenn Greenwald and Ewan MacAskill, NSA Prism program taps in to user data of Apple, Google and others, The Guardian, Published June 7, 2013, https://www.theguardian.com/world/2013/jun/06/ustech-giants-nsa-data. PAGE 11