These days, the businessman Evgueni Prigojine often occupies the headlines as the founder of the paramilitary group Wagner, very active on the Ukrainian front alongside and in competition with the Russian army. Previously, the one who is presented as “Putin’s boss”, because of his past as a restaurateur, made himself known at the head of another army: the Internet Research Agency (IRA), born there ten years in St. Petersburg. The vanguard of a fake news and bot factory suspected of having “trolled” the 2016 US presidential election on social networks on a large scale.
Cook Prigojine’s Recipe: Fake Media, Activist, and Institutional Accounts Making Internet Users Believe that Hillary Clinton Sold Weapons to the Islamic State Organization, That His Campaign Chairman John Podesta Hosted a Pedosatanist Network, or That Pope Francis had supported Donald Trump. The organization, which notably controlled 3,200 Twitter accounts, was specifically charged by American justice in the spring of 2018 for his role in the election. According to the Russian interference investigation report of Special Prosecutor Robert Mueller, she spearheaded “a social media campaign designed to provoke and amplify social discord in the United States.”
To the point of having elected Donald Trump, by tilting the election for a few tens of thousands of ballots in the decisive states of the Great Lakes? A recent study conducted by researchers at New York University (NYU) and published in the renowned journal NatureCommunications seriously temper this tempting accusation. If other research had already nuanced the weight of disinformation on American opinion, its six authors are the first to have been able to analyze in depth the Twitter history of a representative sample of 1,500 American voters during the months leading up to the November 8, 2016 election, at a time when the social network had not yet proceeded to the mass banning of suspicious accounts. It was therefore possible for them to see what information consumed on Twitter those who remained loyal from start to finish to Hillary Clinton or Donald Trump, those who switched from one to the other, those who finally opted for a “small” candidate. or for abstention.
1% of voters have seen 70% of Russian content
They draw from it a series of “concordant evidence” of exposure to Russian disinformation that is less important and more concentrated than is believed, and without tangible effect on the vote. Internet users’ access to the messages of the suspicious Russian accounts, as well as those of other more modest operations carried out by foreign governments (China, Venezuela, Iran), was essentially indirect, that is to say say caused by “retweets” of other accounts more than by a subscription to these accounts. Mostly, phenomenon already observed by other scientific works, In the end, these messages only reached a small number of users: 1% of those surveyed – around 15 people – represent 70% of the cumulative audience of messages from suspicious accounts, and the 10% most affected 98% of the hearing.
These most exposed voters were not the non-aligned or the hesitant of the two camps but on the contrary the most relentless Republicans, those who were already the most inclined to support Donald Trump. In other words, in the great tradition of “selective exposure”, Russian trolls basically preached to the already converted… “Most of the misinformation is about people who are already misinformed, or at least very open to the idea, concludes from his reading of the study Joshua Benton, the founder of the Nieman Lab, a media research organization at Harvard University. If Facebook told you that John Podesta drank children’s blood, chances are you’re already on an, ahem, unusual media diet. A few tweets aren’t going to change your mind about something as meaningful and identifying as the choice of a president.”
These messages also weigh less numerically than conventional political information and communication: for one message from a suspicious Russian account, respondents were exposed to nine messages from elected officials and twenty-five messages from traditional media. On a more qualitative level, this content does not seem to have changed the way respondents represented various controversial subjects, from the extension of health insurance to the wall at the Mexican border through the trade war with China. . Nor did they alter their most essential choice, that of a candidate: NYU researchers identified no visible statistical effect on voting for Donald Trump on November 8, 2016. As the campaigns elections in general, social networks move far fewer votes than we would like to believe.
“Opportunistic real-time chatter”
Of course, the authors note, their work does not exhaust the subject of the influence of social networks on voting, let alone their influence on our political systems in general. Social networks are not limited to Twitter, which has three to four times fewer users in the United States than Facebook. End of 2018, a study by anti-misinformation firm New Knowledge on behalf of the US Senate Intelligence Committee found that the IRA’s operations on the bird social network consisted “essentially of real-time opportunistic chatter”, where the agency had sought to Cultivate your “audience” more thoroughly on Facebook and Instagram. Moreover, the Russian campaigns have not been limited to the creation and dissemination of fake news but have also relied on the circulation of real information, widely taken up by the established media, such as the emails of the Clinton campaign.
If the NYU team warns its readers against the temptation to jump to conclusions too quickly, it is because, from skepticism about the electoral impact of Russian trolls to the non-existence of a Russian destabilization campaign, the shortcut is tempting. The very Trumpist site Breitbart accompanied its summary of the study of the satisfied observation of the collapse of an “old leftist narrative”. Investigative journalist Glenn Greenwald, very critical of the Biden administration and the media presentation of Russian interference, has welcomed the post of NatureCommunications of a “No kidding?”, adding that “Russiagate was, and is, one of the most delusional and deranged conspiracy theories of modern times”. A remark that quickly attracted approval of Twitter’s new owner, billionaire Elon Musk.
Faced with the first reactions to their work, the authors of the study reacted quickly. For one of the co-authors, political scientist Joshua A. Tucker, the Russian trolls were able to succeed even without moving a vote, by reintroducing the worm of doubt into the democratic fruit: by sowing confusion on “the legitimacy of the Trump presidency and the electoral system”, they would have paved the way for voter fraud charges when Joe Biden won. The important thing, in short, would be less to decide the fate of the election than to suggest that we can. To occupy the brains of Americans, failing to make them change their minds.