The history of this volume is in many ways reflective of the topics it tries to cover. As we began assembling the chapters for this book, the 2016 US presidential election controversy was top of mind. When it came to the effect of social media on politics, Russian intervention in the US election served as a wake-up call for internet platforms and as ammunition for their critics. In the wake of that election, governments around the world and the platforms themselves adopted new policies on disinformation, content moderation, and political advertising. As we approached the 2020 election, the companies, the journalists, and the watchdog groups that follow this topic appeared ready to fight the last battle, perhaps with added twists such as “Deep Fakes” or new countries seeking to influence the election outcome.
In the months immediately prior to publication, however, the concerns began to change and to multiply. First, the Covid-19 pandemic eclipsed everything else that was happening in the political world, including what was happening on social media. As people retreated into their homes, they became ever more dependent on technology, even new forms like Zoom and other teleconferencing systems. Social media became a critical means for social connection as physical distancing prevented traditional forms of interaction. If anything, the importance of social media as a source for political news became even greater as people spent more time inside and online.
New concerns about the effect of social media on the information ecosystem likewise emerged, as did new measures taken by the platforms to address them. Medical disinformation spread on social media, with false claims about the origins and spread of the virus and quack cures to address it gaining significant audiences. The platforms responded with dramatic and unprecedented measures: aggressive filtering of content deemed problematic, promotion of content from respected (particularly governmental) sources, and dedicated portions of their websites to assist in providing accurate information and assisting in logistics related to the pandemic response. In many respects, it seemed like the platforms had found a path to redemption from the backlash of 2016, as users began to appreciate the critical function they played in this distressing time, and the fight against medical disinformation (even when it filtered out some “good” speech) is not one that engendered a partisan response.
Just as the new (ab)normal of the pandemic seemed to take hold, however, the killing of George Floyd followed by the protests around the United States dominated the attention of the nation. The history being made in the streets was not a social media story – although publications certainly sought to find the “social media angle” to the protests. To be sure, domestic and foreign sources of disinformation and incitement saw the protests as an opportunity to fan the flames of division. Yet any analysis of the relationship of social media to the protests should not distract from the genuine grievances that led grassroots organizers to take to the streets.
However, at the same time as the protesters were marching, Twitter was marking some of President Trump’s tweets in response as “glorification of violence.” Earlier that same week, it had labeled others, concerning mail balloting, as disinformation, urging users to “Get the Facts” from alternative sources the platform provided. Twitter’s actions led to an unprecedented move by the administration to issue an “Executive Order on Preventing Online Censorship.” The Order asked the Federal Communications Commission (FCC) for a rulemaking on Section 230 of the Communications Decency Act, which is largely credited with creating the free and open Internet by immunizing internet platforms from speech posted by individual users. It called for the FCC to clarify that a platform’s viewpoint-based discrimination could lead it to lose immunity under Section 230. It also called for investigations by the Department of Justice, Federal Trade Commission, and State Attorneys General into the potential viewpoint-based content moderation policies of the platforms. It ended by calling for legislation to implement the positions expressed in the order.
The Executive Order seemed like the first official silo launched from the US government to deal with one dimension of the perceived problems addressed in this volume. However, even as there may be growing consensus on the need for more regulation of social media platforms, there exist significant partisan differences on the perceived problems that social media regulation should address. Some want the platforms to remove more content, whether hate speech, disinformation, incitement, or otherwise. Others worry more about the free speech costs of excessive removal, as well as potential bias on the part of the platform’s content moderators. Although governments around the world have intervened in various ways in each of these domains, the Executive Order was the first attempt in the United States for the government to deal, in a potentially sweeping fashion, with the question of how platforms regulate user-generated content.
As with most government action in this area, the Executive Order was issued without the sound evidentiary basis that analysis of social media data could provide. The folk theory of anti-conservative bias in platform moderation policies comes from a few high-profile cases in which right-wing speakers have had their content or accounts removed; but no one knows how representative these examples are because outsiders do not have access to the necessary data to evaluate whether a politically disparate impact exists in takedowns.
This specter of “legislating in the dark” is not unique to the issue of political bias, however. Indeed, one of the points of this volume – addressed directly in the concluding chapter – is to call for greater access to social media data to better inform legislation concerning disinformation, hate speech, political advertising, and other online content. To craft effective policies, we need public-facing research on the relationship between social media and politics; to carry out this research, access to social media data is paramount. Despite some important steps forward in this regard, it remains the case that the employees of the platforms are the only ones who really know the scale of the problems widely attributed to them. Those of us on the outside must make do with the glimpses provided through publicly available data, which may or may not paint an accurate picture of what is actually going on. As the country is convulsed in ways that, if anything, have made social media even more important for understanding societal development and change, we hope that this volume will serve as a clarion call for the importance of academic access to platform data.
We are grateful for the many individuals and institutions that made this volume possible. The John S. and James L. Knight Foundation provided critical funding for this volume, as well as support for the labs of the two editors and many of the chapter authors. Sam Gill from Knight also provided helpful comments on several chapters. The Social Science Research Council (SSRC) helped organize a conference that generated the idea for this volume, and we appreciate the support of John Ferejohn, Ira Katznelson, and Deborah Yashar, who edit this SSRC series for Cambridge University Press. We also thank our partners from Cambridge University Press, especially Linsey Hague and Raghavi Govindane, who ensured we could get this book published before the 2020 US election. We are especially indebted to the staff at Stanford Law School and the Stanford Project on Democracy and the Internet, particularly Eloise Duvillier and Corissa Paris, who handled all of the logistics necessary to herd the academic cats who wrote chapters for this volume. Finally, we want to thank the chapter authors who contributed their talents to produce what we think is the most comprehensive analysis of this important and timely topic. This book took a long time to produce, but we are extremely grateful for all the work that went into it.
Open access