Nathaniel Persily. Journal of Democracy. Volume 28, Issue 2. April 2017.
If one had tried to write the story of the 2016 digital campaign for the U.S. presidency before knowing the election’s result, the account might have gone as follows. “Hillary Clinton improved on the model built by the successful campaigns of Barack Obama, perfecting the art of microtargeting and the use of online tools to mobilize voters through social media. To be sure, her ability to vastly outspend her opponent on all forms of campaign communication (television, digital, get-out-the vote, or otherwise), as well as her opponent’s relative weakness on all traditional metrics, makes it hard to say whether the new media strategies were decisive. Nevertheless, Clinton’s victory suggests that having a diverse portfolio of media and campaign strategies, while spending an increasing share of campaign funds on digital tools, presages a future in which traditional electioneering becomes married to new technology.”
The actual story of the 2016 digital campaign is, of course, quite different, and we are only beginning to come to grips with what it might mean for campaigns going forward. Whereas the stories of the last two campaigns focused on the use of new tools, most of the 2016 story revolves around the online explosion of campaign-relevant communication from all corners of cyberspace. Fake news, social-media bots (automated accounts that can exist on all types of platforms), and propaganda from inside and outside the United States-alongside revolutionary uses of new media by the winning campaign-combined to upset established paradigms of how to run for president. Indeed, the 2016 campaign broke down all the established distinctions that observers had used to describe campaigns: between insiders and outsiders, earned media and advertising, media and nonmedia, legacy media and new media, news and entertainment, and even foreign and domestic sources of campaign communication. How does one characterize a campaign, for example, in which the chief strategist is also the chairman of a media website (Breitbart) that is the campaign’s chief promoter and whose articles the candidate retweets to tens of millions of his followers, with those tweets then picked up and rebroadcast on cable-television news channels, including one (RT, formerly known as Russia Today) that is funded by a foreign government?
The 2016 election represents the latest chapter in the disintegration of the legacy institutions that had set bounds for U.S. politics in the postwar era. It is tempting (and in many ways correct) to view the Donald Trump campaign as unprecedented in its breaking of established norms of politics. Yet this type of campaign could only be successful because established institutions-especially the mainstream media and political party organizations-had already lost most of their power, both in the United States and around the world.
The void that these eroding institutions left was filled by an unmediated populist nationalism tailor-made for the Internet age. We see it in the rise of the Five Star Movement in Italy, the Pirate Party in Iceland, the “keyboard army” of President Rodrigo Duterte in the Philippines, and the use of social media by India’s Prime Minister Narendra Modi, who has 39 million followers on Facebook and 27 million on Twitter. We see it in the successful use of social media in the Brexit referendum, in which supporters were seven times more numerous than opponents on Twitter and five times more active on Instagram. And we see it in the pervasive fears of European government leaders, who were worried well before the U.S. election that Russian propaganda and other Internet-based strategies could sway their electorates. The Trump campaign, for all its uniqueness, was only the latest to ride a global technological wave that has accompanied deep dissatisfaction with legacy institutions both inside and outside politics.
The “Formal” Digital Campaign
Headlines such as “The Secret Behind Trump’s Comically Bad Digital Campaign?” caricatured the Trump campaign as consisting of little more than large rallies, a Twitter account, and the free media that both engendered. That portrayal has some basis in fact, but it overlooks significant developments that may have future implications for elections at all levels of U.S. government.
Preoccupation with the unique qualities of Trump’s candidacy should not eclipse some of his campaign’s digital innovations. Insiders have suggested that half the campaign’s media budget went to digital media. If that is true, it represents a dramatic shift from previous campaigns. Despite much lower spending overall, the Trump campaign, in fact, spent more on Facebook than did the Clinton campaign. In part, this is due to the campaign’s late start, which forced it to rely on the expertise and personnel offered by the Internet platforms themselves, rather than running most activities in-house. Teams from Google, Facebook, and Twitter were in a single room in the campaign’s digital headquarters in San Antonio, Texas, pitching ideas on how the campaign should spend its money. One of those ideas could be seen on election day itself, as the campaign spent half a million dollars to buy the banner ad viewable on YouTube for the day.
One noticeable difference between the campaigns was Trump’s embrace of Facebook Live. At its inception, the 2016 campaign had been described as the “Meerkat Election,” in reference to the now-defunct live streaming platform. Although that platform died off, live video streaming played an important role in the campaign. For example, Trump provided his own Facebook Live broadcast of the third presidential debate, which included commentary from his surrogates both before and after the event. More than nine-million people saw this broadcast through Trump’s Facebook page, which raised US$9 million in donations. Although the Trump campaign’s digital video presence was generally rough and relied on live events, it lent the campaign an air of authenticity that was lacking in the well-polished, television-quality web videos of the Clinton campaign.
There were three principal components to Trump’s digital campaign operation: the marketing agency Giles-Parscale, the microtargeting firm Cambridge Analytica, and the Republican Party’s digital team. The Trump campaign’s digital director, Brad Parscale, a young Texas businessman with almost no political experience, had performed webpage and other online services for Trump’s businesses. Running the digital operation from a strip mall in San Antonio, Parscale made the important and prescient decision to spend much of his budget on Facebook, which he had used in commercial contexts to target audiences and “lift” brands. Cambridge Analytica, also new to the world of presidential campaigning, had worked on the Brexit campaign as well as the primary campaign of Senator Ted Cruz (R.-Tex.). Employing traditional webbased communication, event promotions, new apps, native advertising (in which web ads are designed to look like articles in the publication containing them), and new uses of social media, the campaign launched 4,000 different ad campaigns and placed 1.4 billion web impressions (meaning ads and other communications visible to individual users).
By Cambridge Analytica’s account, the campaign targeted 13.5 million persuadable voters in sixteen battleground states, discovering the hidden Trump voters, especially in the Midwest, whom the polls had ignored. They also, now infamously, targeted Clinton supporters, especially “white liberals, young women and African Americans,” with communications designed to reduce turnout among those groups. Cambridge Analytica’s psychographic-profiling method, which “scraped” Facebook profiles in order to develop unique voter-targeting models, garnered much attention and criticism, even from Republican campaign consultants. In the end, however, it appears that the firm used more traditional information in order to decide how to direct the Trump campaign’s online advertising. Finally, the campaign benefited from the Republican Party’s data operation, which since 2012 had invested heavily in list-building and other technological tools.
The Decline of the Establishment(s)
Much can be said and will be written about the Trump campaign’s formal data and social-media operation, but the real digital story of the 2016 campaign comes from beyond the campaign organizational chart. In the age of the SuperPAC and partisan media outlets, the formal party organizations and candidates’ campaigns had come to include diffuse networks of groups, consultants, and media companies that specialized in areas such as fundraising, communications, and voter mobilization. Traditional media, which had long played a “gatekeeper” role in setting the agenda for campaign discourse, had been losing ground for some time even before the rise of the Internet. The 2016 contest saw legacy media slip even further. The Trump campaign was in the right place at the right time, with traditional campaign mediators fading badly and no alternative institutions to fill the void.
The Trump campaign’s effective and overwhelming use of Twitter is illuminating in this regard. From August 2015 to election day, more than a billion tweets related to the presidential race appeared on Twitter. By election day, Trump had thirteen-million Twitter followers as compared to ten million for Hillary Clinton. Every tweet from Trump’s account or the account of one of his formal allies would be amplified through retweets from hordes of followers. On average, during one three-week period in mid-2016, Trump’s tweets were retweeted more than three times as often than Hillary Clinton’s, while his Facebook posts were reshared five times more often. This reinforcement went in both directions, as roughly 20 percent of Trump’s own tweets were retweets of the general public, and roughly half his tweets contained links to other news media, as did 78 percent of his Facebook posts.
The power of Trump’s social-media account owed as much to its prominence in legacy media as it did to its propensity to “go viral” online. Whether online, offline, in print, or on the air, Trump set the news agenda. Thanks in part to his followers’ habit of retweeting him, his tweets dominated discussion of the election in every forum. The best social-science study of this admittedly fresh phenomenon finds the following: “Retweets of Trump’s posts are a significant positive predictor of news stories and blog posts … Trump’s tweet volume is a negative predictor of concurrent news coverage … which may imply that he unleashes ‘tweetstorms’ when his coverage is low.” Twitter’s “role is more as a feeding ground for other media, so Trump still relies on other outlets to distribute the content of those tweets to his supporters. In this sense, Twitter is not that different from issuing a press release.” One can see the effect in the tweets from the networks themselves, as during August 2016 the CNN, CBS, MSNBC, and NBC Twitter accounts mentioned Trump almost three times as often as they did Clinton.
To this point, the story of Trump’s social-media dominance is one that reflects a candidate with qualities uniquely tailored to the digital age. Every candidate has assets and liabilities. For Trump, his assets included his fame, following, and skill in navigating the new media landscape. He also figured out that incendiary language could command media attention or shift the narrative. These combined strategies allowed him to garner roughly $2 billion worth of free media during the primaries, and probably a comparable amount during the general-election campaign.
The events of 2016, however, revealed more than just the Internet’s utility for a candidate with Trump’s assets. The Internet reacted and adapted to the introduction of the Trump campaign like an ecosystem welcoming a new and foreign species. His candidacy triggered new strategies and promoted established Internet forces. Some of these (such as the “alt-right”) were moved by ideological affinity, while others sought to profit financially or to further a geopolitical agenda. Those who worry about the implications of the 2016 campaign are left to wonder whether it illustrates the vulnerabilities of democracy in the Internet age, especially when it comes to the integrity of the information voters will access as they choose between candidates.
Given the conflicting definitions of “fake news,” a healthy debate exists concerning its impact on the 2016 election. If the concept includes all false, biased, or objectionable online statements, as some (perhaps even President Trump) would have it, then “fake news” simply becomes a charge to level at a media organization rather than a useful descriptor of a social phenomenon. In evaluating which types of fake news may have electoral impact, though, it may be helpful to differentiate between fake news as satire, fake news for profit, political propaganda, and reckless reporting.
Unlike satirical fake news of the kind presented on television’s Daily Show or online via The Onion, profit-driven fake news purveys falsehoods out of a desire simply to make money. A now-famous group of teenagers in the town of Veles, Macedonia, discovered that publishing pro-Trump and anti-Clinton stories (on close to 140 websites dealing with U.S. politics) could prove to be a profitable venture. As the stories they published grew more outrageous, they would attract more visitors who would, in turn, click through the advertisements appearing on the page. The more clicks they received, the more money these sites would receive through Google’s ad serving process. A story claiming that Hillary Clinton would be indicted over her email server received more than 140,000 shares, reactions, and comments on Facebook. Fake-news purveyors in the United States were able to rake in up to $30,000 per month with stories claiming that an FBI agent had been killed after leaking Clinton’s emails, or that Tom Hanks or Pope Francis had endorsed Donald Trump. Over the campaign’s final three months, the twenty top-performing false election stories generated more engagement than did the twenty top stories featured by mainstream news outlets.
More striking still, the official campaigns would retweet these stories. Donald Trump retweeted one suggesting that his support among blue-collar workers was the highest for any candidate since Franklin Delano Roosevelt. Eric Trump, Kellyanne Conway, and Corey Lewandowski all retweeted an article from a fake-news website (abcnews.com.co) which claimed that Clinton had sent hired protesters to disrupt Trump’s rallies. Indeed, in one of his speeches, Trump even read from what seemed to be a false story that had appeared on Sputnik, a Kremlin propaganda site, about how Clinton confidant Sidney Blumenthal had supposedly judged the 2012 Benghazi attack as a legitimate issue to raise against her. Most infamously, Michael Flynn, Jr., the son of the retired general who would go on to serve briefly as Trump’s national security advisor, tweeted about the Internet-based “Pizzagate” conspiracy theory. That false story was based on conspiracy theories about a child sex ring involving Hillary Clinton and her campaign chairman John Podesta. The rumor proved dangerous, as a man who believed the conspiracy theory entered the Washington, D.C., pizza parlor named in the rumor and fired his rifle there (no one was hurt and the man was arrested).
Propaganda can overlap with satire, profit-seeking fake news, and conspiracy theories, but it involves much more: It is the deliberate use of misinformation to influence attitudes on an issue or toward a candidate. Fake news as propaganda can originate from any node on the diffuse party network and campaign organization described above. It can come from official campaign organs, unofficially allied interest groups, friendly media organizations and websites, foreign actors, or even the candidate himself. In the age of social media, fake news ricochets among these different campaign nodes, moving online and offline as the campaigns, their supporters, and the media repeat stories in the news. The complexity of the network that produces and retransmits fake news often makes it hard to pinpoint the source of a false claim. This is all the more true when the candidate himself retransmits or creates false claims through his social-media account.
We do not yet know how big an effect fake news had on the 2016 campaign. We do know that a majority of Americans have gotten news from social media, with roughly a quarter of respondents saying that they have shared a fake news story. BuzzFeed analyzed six of the most popular hyperpartisan Facebook pages, three on the right (Eagle Rising, Right Wing News, and Freedom Daily), and three on the left (Occupy Democrats, The Other 98%, and Addicting Info). It found that, on the right-wing sites, “38% of all posts were either a mixture of true and false or mostly false, compared to 19% of posts from three hyperpartisan left-wing pages that were either a mixture of true and false or mostly false.” BuzzFeed also found that the “20 top-performing false election stories from hoax sites and hyperpartisan blogs generated 8,711,000 shares, reactions, and comments on Facebook,” which was higher than a comparable set of stories from the top mainstream news outlets.
Economists Hunt Allcott and Matthew Gentzkow question the claim that fake news swayed the election, noting that “even the most widely circulated fake news stories were seen by only a small fraction of Americans.” They observe that television remains the dominant source from which most people in the United States get their news, calculating that “for fake news to have changed the outcome of the election, a single fake article would need to have had the same persuasive effect as 36 television campaign ads.”
Analyzing a database that they compiled of the 156 most-popular fake news stories, Allcott and Gentzkow posit that pro-Trump fake news was shared four times more often than pro-Clinton fake news, with the average fake news headline being seen by 1.2 percent of the U.S. population. The authors surveyed respondents using a mixture of placebo stories (fabricated by the researchers) and actual fake news headlines that appeared on the Internet. The response rate for seeing and believing the fake news (roughly 8 percent) differed little from that elicited by the placebo stories, suggesting that fake news stories by themselves rarely convinced voters of anything.
One can quibble with the assumptions undergirding the above research, but the challenge that it poses confronts anyone trying to estimate the electoral effect of one factor in the complex communications environment of the 2016 campaign. The power of fake news does not derive merely from the changed attitudes of viewers of such stories. It could also demobilize voters by fanning cynicism regarding the candidates and the election. False stories create a blanket of fog that obscures the real news and information communicated by the campaigns. The available academic evidence suggests that viewers have considerable difficulty distinguishing between real and fake news, and that trust in the media is already at an all-time low. The prevalence of false stories online erects barriers to educated political decision making and renders it less likely that voters will choose on the basis of genuine information rather than lies or misleading “spin.”
The power (if any) of fake news is determined by the virality of the lie that it propagates, by the speed with which it is disseminated without timely contradiction, and consequently by how many people receive and believe the falsehood. As with other information or rumors in the offline world, many factors can drive a story’s popularity: its entertainment value, novelty, salaciousness, and the like. But the pace with which lies can travel in the online world is much greater, and different strategies and technologies, such as automated social-media bots, can spread those lies to the right people.
Bots can serve many purposes, some beneficent and others nefarious. They can be used to skew online polls or to write favorable online reviews for restaurants or hotels. They can even be used to automatically generate YouTube videos based on other online content. Of greatest relevance here, bots can spread information or misinformation, and can cause topics to “trend” online through the automated promotion of hashtags, stories, and the like.
During the 2016 campaign, the prevalence of bots in spreading propaganda and fake news appears to have reached new heights. One study found that between 16 September and 21 October 2016, bots produced about a fifth of all tweets related to the upcoming election. Across all three presidential debates, pro-Trump Twitter bots generated about four times as many tweets as pro-Clinton bots. During the final debate in particular, that figure rose to seven times as many.
In many ways, the advent of campaign bots represents the final breakdown in established modes and categories of campaigning. Not only is the identity of the bot generator undisclosed and usually undiscoverable, but even the messages can be automated, reposting other messages and images from the web or automatically responding to posts from the targeted candidate. All the worry about shady outsiders in the campaign-finance system running television ads seems quaint when compared to networks of thousands of bots of uncertain geographic origin creating automated messages designed to malign candidates and misinform voters.
The uncertain geographic origin of bots is also a critical feature that facilitates their use by foreign governments seeking to intervene in another country’s election. We will never know for certain what share of the 2016 campaign’s tweets and social-media posts came from foreign sources, whether these were humans, bots, or “cyborgs” (that is, humans directly orchestrating bots). From the official report of the U.S. intelligence community assessing Russian intervention in the 2016 election, we do know that social media played a key role in Russian strategies to undermine confidence in the election and to magnify stories critical of Hillary Clinton. We also know from earlier reporting that Russia employs teams of people (“trolls”) that post on social media as part of a strategy to influence public opinion. As the report details, “Russia used trolls as well as RT as part of its influence efforts to denigrate Secretary Clinton.” Specifically, the report states that “a journalist who is a leading expert … claimed that some social media accounts that appear to be tied to Russia’s professional trolls-because they previously were devoted to supporting Russian actions in Ukraine-started to advocate for President-elect Trump as early as December 2015.”
The Russian campaign of hacking and propaganda employed a range of social media, webpages, and radio and cable television stations. Sputnik and RT provided a social-media and web presence on all relevant platforms, as well as on television and radio, typifying the fluid media environment that characterized the campaign in general. The lines between insiders and outsiders, legacy and “new” media, or media and campaigns became increasingly blurred. The Russian propaganda campaign used “Russian Government agencies, state-funded media, thirdparty intermediaries, and paid social media users or ‘trolls.'” As with Internet-based campaigns in general, its effect was magnified by the assistance, whether deliberate or unwitting, of those inside and outside the official campaigns who retweeted, reposted, or otherwise spread messages crafted elsewhere.
A New Playbook, a Novel Challenge
If the 2008 and 2012 U.S. presidential campaigns had seemed to confirm Internet utopians’ belief that digital tools enhance democracy by expanding citizen empowerment and engagement, the 2016 campaign highlighted the challenges that the Internet poses for American democracy, and perhaps democracy in general. The surprising robustness of the campaign mounted by Bernie Sanders, the Independent senator from Vermont who challenged Hillary Clinton for the 2016 Democratic nomination, seemed to pick up where Obama’s two campaigns and even Howard Dean’s in 2004 had left off: A candidate running against the establishment proved able to raise money, organize supporters, and mobilize voters as never before. Trump fulfilled this promise too: He showed how the Internet can enable an outsider to run for-and win-the presidency by means of a nontraditional campaign despite being outspent two-to-one by an establishment opponent.
From the point of view of the health of liberal democracy, the Internet’s great promises are also its pitfalls. Its liberating, anti-establishment potential can be harnessed by demagogues who appeal to the worst impulses of the mob. By aiding and abetting the disruption of established (and in some ways, outdated) institutions, such as political parties and the media, the Internet left a void that could then be filled not only by direct appeals from candidates, but also by fake news and propaganda. Furthermore, the anonymity and lack of accountability that give Internet speech its power-in whistleblower cases or in repressive contexts such as those faced by Arab Spring demonstrators-also enable foreign powers to intervene secretly in campaigns and allow trolls to commit racial and sexual harassment. Finally, the Internet’s unprecedented ability to facilitate the targeted delivery of relevant information, marketing, and even friendship also leads to the bubbles, filters, and echo chambers that shelter people from information that might challenge the messages sent to them by campaigns, partisan media, or social networks.
What the success of the Trump campaign demonstrated is that virality is now the coin of the campaign realm. Those candidates and strategies that can generate “shares,” “likes,” and “followers” have an advantage over those that cannot. Of course, enthusiasm and popularity are not campaign assets novel to the Internet age, but what the Internet uniquely privileges above all else is the type of campaign message that appeals to outrage or otherwise grabs attention.
The politics of never-ending spectacles cannot be healthy for a democracy. Nor can a porousness to outside influences that undercuts the sovereignty of a nation’s elections. Democracy depends on both the ability and the will of voters to base their political judgments on facts, or at least on strong intermediary institutions that can act as guardrails to channel decision making within the broad range of democratic alternatives.
The premium placed on virality of messages, the threat to accountability posed by unrestrained anonymity, and the undercutting of sovereignty presented by an open Internet pose novel challenges for democracy in the United States. The election of Donald Trump may, indeed, be a “one-off,” as it is difficult to think of many people with his particular personality qualities, strengths, and motivations. Nonetheless, the playbook for one type of successful candidacy and campaign in the Internet age has now been demonstrated. Whether others can succeed with the same playbook remains to be seen.
Epilogue: Post-Election Reforms
Following the 2016 election, several Internet platforms changed their policies concerning information on their sites to address perceived shortcomings of the communications environment. Google, Facebook, and Twitter each enacted new rules for news and other communication on their platforms, based on complaints related to the 2016 presidential campaign. The nature of these attempts, however, says much about the novelty of the challenges that these platforms face and the intractability of problems such as fake news and hate speech in the Internet age.
Google and Facebook attacked the easiest problem first, and did so within two weeks following the election. They both adopted policies at-tempting to target fake-news-for-profit. They tried to remove the economic incentives that they had created for those sites to drive traffic based on outrageous, clickbait headlines. In particular, Google now bars certain fake-news sites from its advertising network (AdSense), meaning that such sites will not be able to earn money from having Google place an ad on their site. The regulated sites are ones that Google says “misrepresent, misstate, or conceal information about the publisher, the publisher’s content, or the primary purpose” of the site. Facebook took similar steps with changes to its Audience Network Policy, to try to drain support for the most egregious sites that simply make up stories for profit.
While initially rejecting the notion that fake news on its platform swayed election results, Facebook quickly recognized that it needed to do something to assuage widespread concerns that it had been responsible for misinformation during the campaign. At the same time, it had only recently emerged from the controversy concerning the editing of its trending news feature, when it came under attack for sublimating conservative news stories. Then as now, the company was caught between being the principal promoter of fake news and the chief policeman of the truth. To thread that needle, Facebook now outsources fact-checking to established organizations (such as Snopes, Factcheck.org, ABC News, AP, and Politifact), but relies on its users to flag news stories as fake to trigger the fact-checking process. If two of those organizations consider the story false, the news item is not removed, it is simply flagged in the newsfeed as “Disputed by 3rd Party Factcheckers,” with a link to the article that disputes it. Those who would forward such stories are also warned before doing so that they would be forwarding a disputed story. Needless to say, however, the damage of a popular fake-news story is usually well done by the time it is fact-checked and flagged.
Twitter, for its part, took steps after the election to address hate speech on its platform. The effective use of Twitter by the alt-right and hate groups, during an election campaign when race, gender, and religious issues were prominent points of debate, led to the perception, at least, that the platform was uniquely enabling of online bigotry and harassment. Political scientists who study online abuse have not unearthed a recent increase in hate tweets, even if journalists, in particular, were increasingly targeted during the campaign. Nevertheless, just as the impact of fake news might be indeterminate but deserving of attention, so too Twitter considered hateful content and conduct to require new rules. Following the election, Twitter expanded its hateful-conduct policy, began to allow users to hide tweets with certain words (even beyond obvious racial epithets), and made it easier for persons other than those targeted by a tweet to report online abuse.
Twitter, Facebook, and Google acknowledge that these initial steps will not be the last ones taken to address the criticisms relating to their involvement in the 2016 campaign. With every move to restrict election communication and advertising on their sites, however, they recognize that they trade one set of criticisms for another. Try as they might not to be media companies, they nevertheless have power far in excess of that which legacy media institutions had in their heyday, let alone today. Especially in an environment in which the regulated speech-whether hate speech, fake news, or otherwise-tends to predominate on one side of the political spectrum, they cannot escape the charge that their new rules are biased either in intent or in effect.
Perhaps the more intractable problem these companies face is that their business models naturally hinder the kind of approaches necessary to combat the communication pathologies identified in this campaign. The strength of a search engine like Google comes from the relevance of the search results and other information it serves to its users. One man’s relevant result, however, is another’s filter bubble, and the search for campaign information will naturally lead users down paths trodden by their prior ideological commitments. (Indeed, even the task of separating out political or campaign-relevant information for special treatment poses a problem, as all information, let alone entertainment, forms the contours of the bubble in which political news is experienced.) Similarly, Facebook sets as its mission providing the most “engaging” and “meaningful” experience for its users. The “search for truth” is necessarily far down the list of priorities for the social network, just as it is for its users, who will often find false, negative, bigoted, or other outrageous speech to be more meaningful and engaging. Finally, Twitter prides itself on the anonymity it guarantees its users, which enables subversive speech against dictatorial regimes just as it facilitates tweetstorms of hate and threats. Like the other platforms, it, too, places a premium on popularity, rather than some other measure of merit, in sifting through and serving information to its users. Popularity, of course, only occasionally derives from sources that might further responsible democratic political decision making, and more often is earned through appeals to emotion, interest, and fear.
Reluctantly or not, these platforms are the new intermediary institutions for our present politics. The traditional organizations of political parties and the legacy media will not reemerge for the Internet age in anything like their prior incarnations. Unlike those earlier institutions though, these new firms were not created principally to serve democratic values and do not have as their lodestar the fostering of a well-informed and civically minded electorate. It would be easy to fault them for this, but to some extent, they have stumbled into the darkness of politics and cannot find a clear path out.
Social networks and search engines serve as tools for all aspects of life, not merely as context-specific instruments for politicians or voters. The intersection of these tools with the democratic process is critically important, but a small part of what they do. With the deterioration in democratic values occurring both on- and offline, we should not expect technology to rescue us from the historical and sociological forces currently threatening democracy, even if that same technology facilitated the disruption in democratic governance in the first instance.