Is Social Media a Threat to Democracy? Looking Forward Social media platforms are ingrained in our daily lives and provide much of the infrastructure of democratic debate. They have essentially become the modern “public square,” and they have command over both our attention and much of our personal data. As we study this issue more closely, it’s clear that we have much more to learn and many difficult questions to answer. Among these questions: what responsibility do these companies have in acknowledging their role and taking accountability for shaping information and news, and affecting public opinion and behavior?40 Some argue that just as it would be unconscionable for the editor of a traditional media outlet to not care about what he or she publishes, promotes, or endorses, the designer of a social media site’s non-neutral algorithms has a responsibility to acknowledge his or her code’s role in selecting which content to prioritize or de-prioritize, and the implications that these models have on democratic deliberation. The Omidyar Group’s Near Term Approach The Omidyar Group’s approach to the challenges of online hate speech includes a partnership in the U.S. with the Anti-Defamation League to establish a Silicon Valley center that will focus on finding solutions. Indeed, platforms have also struggled to remove content judged to be problematic for several reasons; given the difficulty of setting up normative codes, content standards, and policing capacity that would allow for different outcomes. Such efforts— whether for terrorism, bullying and violent language, or targeted trolling and hate—create challenges for platforms that are reluctant to determine who adjudicates speech, where to draw the line regarding freedom of expression, and how to differentiate legitimate dissent from calls to violence or bullying. Of late, these platforms have formed a shared database to address content that violates their policies. Additionally, to what extent can we expect these platforms to internally “self-regulate” their content and activity, especially if it has implications for their revenue model? While increased oversight might strengthen long-term trust and engagement, it might also come with nearer-term attention costs that would require mitigation. To an extent, firms like Facebook and Google have announced steps to combat these issues themselves—e.g. Google has founded First Draft News Coalition, and Facebook is joining The Ford Foundation, Mozilla, and Craigslist Founder Craig Newmark in funding the News Integrity Initiative.41 While seen as positive, these steps do not fully answer questions about their commitment.42 Mathew Ingram, Mark Zuckerberg Finally Admits Facebook Is a Media Company, Fortune, Published December 23, 2016, http://fortune. com/2016/12/23/zuckerberg-media-company/ 41 Democracy Fund works closely with the News Integrity Initiative 42 Have the platforms determined they need to make substantial investments to be more than agnostic about their impact on 40 © Copyright The Omidyar Group 2017. All rights reserved. Other examples of self-regulation might include: • Technological/engineering decisions taken by social media firms to change how their algorithms work to address these issues • Decisions to allocate at-scale resources to community standards, including effective grievance redressal mechanisms •  Adherence to self-generated industry standards and commitments to minimizing the unintended consequences of social media technologies and platforms •  In the interim, measures such as partnerships with institutions to seed and innovate solutions on critical topics, with which the social media platforms acknowledge they require assistance and external expertise, e.g. Facebook’s decision to share • Information and partnership with the International Fact Checking Network and with Germany’s Correctiv to combat the spread of mis-and dis-information • Social media platforms could also commit to greater transparency around dark advertising, collaborating with researchers and civil society to provide access to data about how information is being shared and targeted, especially when this is shown to be in the public interest. democracy, and ensure the platform is as pro-democracy as possible? Will they be willing to bear the at-scale costs of doing so and engineer solutions to create disincentives for viral deceptions, less homogeneous networks of thought/echo chambers, etc. at the cost of making the platform less “addictive”? And what can the public do to encourage such efforts? PAGE 14