Troll Factories

Donald Trump’s​ press conference on 5 November, in which he claimed he had won the election and said there had been mass voter fraud, was like a scene from a dystopian movie. His allegations were so egregious that TV networks cut away from their live broadcast of his remarks. Yet in the months and years leading up to that moment it was as if no lessons had been learned from Trump’s victory in 2016, or from the Brexit campaign in the UK earlier that year. Populist leaders in Western democracies have continued their demagoguery, and mainstream media outlets have continued to indulge them. Meanwhile, citizens have lost faith in the media’s capacity, or willingness, to tell them the truth. Those on the left accuse the press and broadcasters of indulging the right; those on the right accuse the media of not indulging them enough.

Much of the blame, as the liberal consensus has it, lies with the misinformation emanating from Russia and circulating on social media. Russia is pro-Trump; it wants to disrupt elections and break up the EU; it causes chaos, interferes with ballots. We are told that an ‘information war’ is being waged by authoritarian states against Western democracies. All this is to ignore the fact that misinformation networks don’t respect state borders or distinctions between political systems. The ‘Russia Report’ published this summer by the UK Parliament’s Intelligence and Security Committee, while rightly cautioning against exaggerations of the Kremlin’s foreign policy ambitions and its capacity to fulfil them, nevertheless falls into the trap of subscribing to the ‘information war’ paradigm. It argues that ‘the operational role’ in protecting our ‘democratic discourse’ must ‘sit primarily with MI5’, the UK’s counterintelligence and security agency. This idea – that public discourse should be protected by intelligence services – is something we might ordinarily expect from the Kremlin, not from a Western government. In this case, it stems from the assumption that the main threats to democracy come from without.

It’s easy to see how the ‘Russia factor’ came to play such a dominant role. After the US elections in 2016, an investigation by American intelligence agencies found that from 2013 onwards a Russian government-affiliated troll factory called the ‘Internet Research Agency’ had been attempting to influence US voters by means of an extensive, well co-ordinated campaign of astroturfing – the use of online fake accounts to mimic grassroots activities. Thousands of fake accounts purportedly belonging to US citizens or organisations were set up by the IRA, and began to disseminate content across social media platforms. In the weeks leading up to the election, these accounts focused increasingly on circulating pro-Trump propaganda. Alongside the IRA campaign, Russia-affiliated hackers infiltrated the computer networks of the Democratic National Committee and leaked into the public domain emails that embroiled Hillary Clinton in scandal. This wasn’t the first attempt by a foreign state to interfere in elections, but it was the largest known operation of its kind against a Western democracy. And so the actions of the US’s Cold War adversary came to be seen as a means of explaining away the Trump phenomenon.

Many politicians, journalists and observers were quick to assert that the activities of the Russian troll factory were a ‘significant’ or even ‘decisive’ factor in Trump’s victory in 2016 – see the New Yorker (issues passim) or Nina Jankowicz’s recent book How to Lose the Information War: Russia, Fake News and the Future of Conflict (Bloomsbury, £20) – and the threat of Russian interference was a continual source of anxiety in the early stages of this year’s campaign. But the focus should have been on sources of right-wing misinformation in the US itself, which have been the subject of several recent academic studies. Researchers at Indiana University working on the role of astroturfing in US elections detected the use of false accounts in support of Republican candidates in the midterm elections in 2010. In 2017, a report by Oxford University’s Computational Propaganda Research Project suggested a high density within the US of ‘cyber troops’: ‘government, military or political party teams committed to manipulating public opinion over social media’.

The proliferation of misleading allegations of mass fraud in this year’s presidential election makes clear how counterproductive it was to view Trump’s victory in 2016 as a consequence of the threat to democracy posed by external authoritarian actors. Right-wing conspiracy theories are spread by homegrown US outfits such as Project Veritas, started in 2010 by James O’Keefe, and Infowars, set up a decade earlier by Alex Jones. RT, the Russian state’s international broadcaster, also relays misinformation, but it has far smaller audiences in Western societies than local providers of hyper-partisan news. The coverage of Rudy Giuliani’s accusations of electoral fraud by Newsmax, a conservative media organisation that has been operating out of Florida for more than twenty years, had 1.6 million views within hours of being posted on YouTube on 10 November. By comparison, RT America’s coverage had nine thousand views by the end of the following day. Newsmax, Project Veritas and Infowars are followed by millions, leaving the online presence of broadcasters such as CNN, ABC and the BBC in the shade.

It is only in the last month that mainstream media outlets have begun to pay more attention to American misinformation sources, scrutinising the role of sites like Infowars and new platforms such as Parler in enlisting supporters at pro-Trump rallies. Trump and Trumpism is not a glitch in American democracy that starts and ends with the man himself. In How Democracies Die (2018), Steven Levitsky and Daniel Ziblatt’s analysis helps us to understand why more than seventy million US citizens voted for Trump in November, and why multiple Republican officials supported claims of electoral fraud despite the absence of evidence. For a democratic system to endure, trust in the integrity of elections is essential. Levitsky and Ziblatt show that the casting of doubt on the electoral process, particularly by Republican officials, began in the US at the end of the 1990s, implicitly delegitimising the Democrats as popular representatives. Even before the 2016 vote, polls showed that the majority of registered Republicans believed that significant levels of fraud occurred during US elections, although by any real measure electoral irregularities in the US are extremely rare.

Unfortunately, the reaction of the press and the Democratic Party to evidence of Russian interference in 2016 only contributed to the erosion of public confidence in elections. The New York Times created a tag – ‘Russian Hacking and Influence in the US Election’ – under which to gather its many articles on the theme. There is no doubt that Russia did try to interfere in 2016, but evidence of Russian intentions has been systematically misread as evidence of Russian success in actually changing votes. We now have a body of research showing that the impact of the IRA on US voting in 2016 was minimal. The circulation of tweets and other social media posts can easily be measured, and they didn’t swing the election. Actual hacking, on the other hand, did have an effect – but only because of the way it was reported in the media, undermining confidence in Hillary Clinton and the Democratic Party.

Those researching the role of the IRA have noted a curious fact: the surprisingly low level of Russian troll activity in the swing states that decide the outcomes of US elections. The most active IRA accounts in 2016, with the largest number of followers, purported to belong to a local Republican Party organisation in Tennessee – a solidly red state in the previous four elections. Studies of the operation of troll factories in various countries suggest an explanation: it’s a result of the way the employees of such outfits are remunerated. Since the pay of rank and file trolls isn’t linked to outcomes but to their success in circulating information, it doesn’t matter to them whether their activities actually change voters’ preferences. In order to gain more followers, the trolls follow one another’s accounts, retweet one another’s posts and use identical content. The practices that maximise their wages necessarily reduce the political effectiveness of their activities. It made financial sense for IRA employees to disseminate right-wing posts to friendly audiences in Tennessee rather than spend time and energy on posts that might or might not have resonated in swing states.

Despite being an operational failure, the IRA was a strategic success, since the airtime devoted to it – all the while exaggerating its actual effects – cast further doubt on the integrity of electoral processes in the digital age. Trump and his supporters can now spin to their advantage the perception created in 2016 by those opposed to him that the electoral process can be successfully rigged. The representation of Russia as the cause of electoral swings towards politicians who violate democratic norms serves the Kremlin’s interests, too, attributing to Putin far greater power than he actually has. It also assists the Russian political elite in arguing that the West is paranoid and Russophobic. The same applies to talk of the information wars being waged by any foreign state – China, North Korea, Iran. The picture of misinformation as an external threat to be countered by a democratic fourth estate, with the support of the intelligence and security services, deflects us from confronting the wider pathologies in media operations at home. The fallout from this year’s US presidential election shows that this confrontation can no longer be postponed.