Messing with the Enemy

by Clint Watts

  7

  Postmortem

  The postmortem on Russia’s influence and meddling in the presidential election of 2016 began well over a year ago and may never end. Less than a month after the election, social media influence became a fixation with journalists as they looked for an explanation as to how Donald Trump had beaten all the odds. He was completely unconventional, uninformed, unlikable in so many ways, and yet he had become the leader of the free world. Fake news entered the American lexicon, and my team’s pre-election detailing of Russian active measures on the internet was now the subject of hot debate. Had fake news swayed the U.S. presidential election? A Washington Post article cited our study, and soon left-leaning trolls, led by the self-righteous Glenn Greenwald, of the Intercept, and the always bitter Matt Taibbi, of Rolling Stone, grouped me with McCarthyites seeking to suppress free speech. Right-wing fanatics, aided by Russia’s trolls, presented me as a Clinton apologist, a sore loser always out to get President-elect Trump. I’m certain I’m neither and also can’t be both at the same time.

  Beyond the public back-and-forth, social media companies began digging into the data, and what they found spelled dangerous trends for democracy. Americans were increasingly getting their news and information from social media rather than mainstream media. Users weren’t consuming factual content. Fake news—false or misleading stories from outlets of uncertain credibility—was being read far more than that from traditional newsrooms. BuzzFeed News analysis of the final three months of the campaign showed that Facebook users accessed false news stories at rates higher than mainstream news. EndTheFed.com and Political Insider produced four of the five most read false news stories in the three months leading up to the election. One story famously and falsely claimed that Pope Francis had endorsed Donald Trump and another story claimed that Hillary Clinton’s emails hosted on WikiLeaks certified her as an ISIS supporter. Three of the five most read articles, though false, promoted and referenced Russia’s hacking of Americans and delivery of stolen contents to WikiLeaks.1 Throughout December, fears of Russian election manipulation grew, and each day brought more inquiries into how Russia had trolled for Trump.

  “There’s no evidence of collusion” had become a constant scream from Trump supporters by the summer of 2017. Trump opponents, self-labeled on Twitter as the #Resistance, saw conspiracy and Russian collusion at every turn, hyperventilating with each new revelation connecting Team Trump with Russia. Months of Russia investigation by Congress, the firing of FBI director Comey, and the appointment of Special Counsel Robert Mueller, a former FBI director, to examine Russia’s interference in the election have extended Vladimir Putin’s victory over the United States. The American electorate remains divided, government operations are severely disrupted, and faith in elected leaders continues to fall. Americans still don’t grasp the information war Russia perpetrated against the West, why it works, and why it continues.

  The Russians didn’t have to hack election machines; they hacked American minds. The Kremlin didn’t change votes; it won them, helping tear down its less-preferred candidate, Hillary Clinton, to promote one who shares their worldview, Donald Trump. Russia’s ability to influence the 2016 vote comes from many structural factors in American democracy. Narrow bipolar races make foreign influence such as that by the Russians particularly easy and effective. In the cases of both the United Kingdom’s 2016 Brexit vote and the U.S. presidential election, the difference between the two sides amounted to merely one percent. A slight nudge at the polls provided by strategically leaked kompromat can send a preferred candidate over the top or suppress turnout for supporters of Kremlin foes. In two of the past five elections, the candidate who lost the majority vote—Bush in 2000 and Trump in 2016—ascended to the presidency. Electoral College disenfranchisement gives outside influencers the ability to enter into an electoral space and strategically target voters in only a few states with propaganda. From a laptop, anyone can microtarget a state’s or county’s voters on social media through deliberate analysis. The duration of U.S. presidential campaigns also assists adversary manipulation. Trump’s rise occurred over almost sixteen months, stretching from the summer of 2015 to the end of 2016. Long primary and election campaign seasons provide extended periods for Russian propagandists to infiltrate key audiences and then direct them on strategic themes and messages supportive of Kremlin-preferred candidates.

  Americans’ rapid social media consumption of news creates a national vulnerability for foreign influence. Even further, the percentage of American adults fifty and older utilizing social media sites is one of the highest in the world, at 50 percent. Younger Americans, aged eighteen to thirty-four, sustain a utilization rate above 80 percent.2 Deeper analysis by the Pew Research Center shows that U.S. online news consumers still get their information from news organizations more than from their friends, but they believe that the friends they stay in touch with on social media applications provide information that is just as relevant. Online news consumption moves increasingly each year from desktops and laptops to mobile devices; mobile has seen more than 20 percent growth in just three years. Social media apps such as Facebook dominate user time on mobile devices, pointing to the natural shift of news consumption to these applications. The American move to online and mobile has been accompanied by a massive growth in online news outlets that suit the preferences and biases of Americans. Diffusion enables Russian active measures to infiltrate American audiences through multiple strategic placements narrowly targeted to audiences that can tip an election.

  The Columbia Journalism Review analyzed the news outlets most frequently shared by supporters of Trump and Clinton. Fans of both candidates demonstrated a proclivity for outlets supporting their political biases, but the differences between the two camps were stark. On social media, Clinton supporters shared the Washington Post, Huffington Post, and New York Times the most. Trump supporters far and away preferred Breitbart, the Hill, and Fox News. Further down the list, Clinton supporters gravitated to a wide range of liberal outlets, most fairly well known. Trump’s camp, though, included a long list of lesser-known outlets, including the controversial Infowars, a media organization known for denying the occurrence of the Sandy Hook shootings and one I’d witnessed routinely regurgitating Russian propaganda.

  A cursory look at the Columbia Journalism Review’s media map demonstrates how social media encourages information bubbles for each political leaning. Conservatives strongly centered their consumption around Breitbart and Fox News, while liberals relied on a more diverse spread of left-leaning outlets. For a foreign influence operation like the one the Russians ran against the United States, the highly concentrated right-wing social media landscape is an immediate, ripe target for injecting themes and messages. The American left shows to be multipolar, littered with fringe outlets and causes, making concentrated foreign influence more challenging; spreading the Kremlin message thus requires influencing many outlets rather than one or two. Above all, the study shows how damaging Clinton’s emails were to the campaign. Even among mainstream news outlets, Clinton’s emails, a topic that the Kremlin pushed massively, was the subject of more than 65,000 sentences—nearly double any other campaign issue and far outweighing any Trump scandal.3

  Our pre-election article on Russian social media active measures placed me squarely in discussions regarding whether President Trump legitimately won the election or if Russia did it for him. These debates rage on today, well into Special Counsel Robert Mueller’s investigation and multiple Senate and House intelligence committee investigations into Russian meddling. Did Russia win the election? Did Russia collude with the Trump campaign?

  These questions are not mutually exclusive. They overlap and diverge, and the biggest ones will prove vexing and, ultimately, indeterminate.

  In my opinion, Russia absolutely influenced the U.S. presidential election of 2016. The single largest theme echoed throughout the campaign was Hillary Clinton’s emails. Russia’s theft and repeated rele
ase of emails from the DNC and the Clinton team powered and sustained a narrative of corruption, criminality, and conspiracy that clouded the Clinton campaign from start to finish. Trump and his advisers directly cited WikiLeaks releases and opportunistically repeated Russian narratives. Chants of “Lock her up!” and “emails!” still—more than a year after the election—ring out at Trump rallies. Some have suggested that political campaign propagandists were the ones using social bots to influence Trump supporters, but these claims fail to account for the value of the material that Russia obtained through hacking, which powered those political attacks against Clinton. Same goes for fake news peddlers in Macedonia and other locales who used sensational headlines and stories to create clickbait for advertising revenue. Of course, they influenced the election, but their bogus narratives benefited from the one thing Russia did that no one else could do: strategically hack and release kompromat.

  Russia didn’t need anyone’s help to target Americans online; instead, it helped everyone else by hacking and releasing confidential secrets. Some conspiracies claim that Russian social media influence efforts worked in concert with the Trump campaign to microtarget specific portions of the electorate. I’ve seen no evidence to support this theory, and while I do believe that political groups benefited from Russian influence, that doesn’t necessarily mean that those political groups knowlingly colluded with Russia in cyber influence operations. I observed Russian social media operations dating back to 2014. Many of these accounts ultimately turned toward U.S. audiences in 2015 and later informed online discussions of the presidential election in 2016, but they predated the Trump campaign. Political social media advertising can easily and quickly repeat effective social media manipulation from Russia or other influencers without directly coordinating their efforts.

  Russia alone did not win the election for Trump. It certainly helped the race remain close at times when Trump might have fallen completely out of the running. A prime example of its assistance is the strategic dumping of the stolen John Podesta emails less than an hour after Trump’s disastrous, sexist “grab ’em by the pussy” comments hit the airwaves. Efforts like this offset attention from potentially catastrophic incidents in the Trump campaign. But Russian support for Trump and derision of Clinton can’t be separated from the FBI investigation of then–Secretary of State Hillary Clinton’s email server, the July 5 announcement by FBI director Comey closing the case, and the monumental shift arising from Director Comey’s announcement, one week before the election, that new emails had surfaced, requiring a reopening of the investigation.

  Clinton, Putin, and Trump were all equally shocked that Trump won the election. Without the Comey letter, I believe Clinton would have won the election. Russian influence networks also shifted their themes to voter fraud and election rigging in preparation for a Clinton win, as a way to undermine her mandate to govern should she lose. The dossier compiled by ex-MI6 agent Christopher Steele on Trump, and a study published by the Russian Institute for Strategic Studies, a Moscow-based think tank established by Boris Yeltsin in 1992, both pointed to this shift in strategy by October 2016.4 Without the Russian influence effort, I believe Trump would not have even been within striking distance of Clinton on Election Day. Russian influence, the Clinton email investigation, and luck brought Trump a victory—all of these forces combined.

  Facebook, Google, and Twitter are likely the only ones with sufficient data to determine whether Russian influence won the election for Trump. They could track the spread of specific Russian-powered themes to specific geographies in the United States. Clicks in key swing states could be calculated and compared with political polling to see how certain themes or messages emanating from the Kremlin may have influenced perceptions. The equation would be fairly straightforward: How did false news and manipulated truths from Kremlin influence networks shift votes from Democrats and independents to Trump or bring about decreased voter turnout among Democrats in swing states?

  Even if this electoral arithmetic were possible, one metric remains elusive: accurate political polling going into Election Day. Almost every poll leading up to the U.S. presidential election was wrong, and pollsters missed the mark repeatedly during the 2016 campaign year. Voter shifts due to Russian hacks and influence operations will be impossible to accurately account for, as there is no true measure for public opinion. No one will ever be able to prove without a doubt whether Russia did or did not win the election for Donald Trump.

  * * *

  Having watched Russian influence leading up to the election, my hypothesis is that Putin won at least two states for Trump: Michigan and Wisconsin. Here’s why. The first step when analyzing influence is to look at the audience’s media usage. Of the key swing states in the 2016 presidential election, Wisconsin and Michigan have slightly higher internet penetration and mobile usage than the others, so there’s a greater chance that voters there were exposed to Kremlin influence.5 Next, Michigan and Wisconsin proved to be the closest races in the election, decided by less than 1 percent of the votes. Only 10,700 votes in Michigan and 22,700 votes in Wisconsin separated the two candidates.6 Prior to the general election, Hillary Clinton struggled in these two states, losing both primaries to Bernie Sanders. These losses made Michigan and Wisconsin voters ripe for all three of the principal themes Russia pushed leading up to the election: Clinton’s emails, her corruption and potentially poor health, and narratives of Bernie Sanders getting a raw deal from the Democratic National Committee. A minor theme pushed by Russia’s social media operations sought to encourage Jill Stein supporters to make it to the polls, even though she had no chance of winning. On Election Day, Democratic turnout for Hillary Clinton in both states was lower than in previous elections. Russia’s pump for Trump, derision of Clinton, advocacy for Sanders, and slight nudging for Stein could easily account for this small turnout differential.

  “But I’m not on social media, so I wasn’t influenced by the Russians” consistently rates as a top counter to ideas of Kremlin shaping of American minds. These voters mistakenly believe that the social media and mainstream media worlds don’t cross, but major news networks increasingly rely on social media to generate their stories. After the election, one set of researchers at the University of Indiana undertook the due diligence of analyzing the relationship between social bots, fake news, mainstream media, and influence.7 After analyzing fourteen million messages spreading four thousand claims on Twitter from before and after the 2016 election, the Indiana team discovered that social bots provide an essential amplification effect, making false claims go viral across Twitter. These bots also directly targeted influential users on social media, such as political candidates and the media reporters and producers who cover these candidates. Automated social bots influence what mainstream journalists report on.

  Whether a voter uses Twitter or not, many of the stories one consumes in print, radio, and television media originate on social media. Not using Twitter does not prevent one from being exposed to what goes viral on Twitter. Any casual television watcher or radio listener has been exposed to the endless debates on President Trump’s Twitter usage since assuming office. Even today, the president’s Twitter dialogue may be the largest topic of discussion on many mainstream outlets. No border separates social and mainstream media—the two are symbiotic and synergistic.

  Social media companies came under fire from Congress and the American public in 2017 for not detecting the Kremlin’s manipulative measures on their platforms. Facebook seemed to be the first to catch on to the shenanigans coming from the St. Petersburg–based Internet Research Agency. In response, the platform shuttered thousands of accounts prior to the 2017 French presidential election. In October 2017 the social media giant revealed that more than a hundred confirmed Russian-backed pages had spent more than $100,000 on ads. Facebook’s general counsel, Colin Stretch, described these ads as “deeply disturbing . . . an insidious attempt to drive people apart” using race, religion, gun rights, and gay and tr
ansgender issues to inflame social and political divides. These known Russian-backed accounts posted more than 80,000 pieces of content and may have touched up to 126 million users.8 Fake Russian social media accounts created 129 Facebook events, which reached 300,000 users and gained 62,500 users saying they’d attend and another 25,800 users expressing interest in the events. One event page, called Heart of Texas, called for the secession of Texas from the United States. It initiated a “Stop Islamization of Texas” event in Houston, trying to rally protesters at an Islamic center on May 21, 2016. Then the trolls hosted a United Muslims of America page to draw people to another Houston event called “Save Islamic Knowledge,” occurring at the same time. The Kremlin used Facebook to pit competing factions against each other at a time and place of its choosing, and it attempted this among nearly all American divides around the country.9

  Google found Russian influence at a smaller level than Facebook, but the manipulation was equally disturbing. Internet Research Agency accounts bought $4,700 in advertising and, through eighteen channels, hosted more than 1,000 videos receiving more than 300,000 views.10 The depths of the Kremlin’s efforts were discovered when it was revealed that it had created a YouTube page called Williams and Kalvin. The page’s videos showcase two black video bloggers, with African accents, appearing to read a script stating that Barack Obama created police brutality and calling Hillary Clinton an “old racist bitch.” Some dismissed the effort as a phony failure, but the Williams and Kalvin Facebook page garnered 48,000 fans.11 Russian influence operators employed most every platform—Instagram, Tumblr, even Pokémon Go—but it was the Kremlin’s manipulation via Twitter that proved the most troubling.