In a finding that has implications for the 2022 midterm election, Cornell researchers found that Russia attempted to distract liberal voters during the 2016 presidential campaign with a seemingly innocent weapon — tweets about music and videos — that took a page out of its domestic disinformation -Take playbook.
The strategy is similar to techniques used by autocratic governments that control their national media, such as Russia and China, which “flood” social media with entertainment content to distract their citizens from national events, such as protests, that they do not wish to cover.
“We have seen flooding as a social media strategy in both China and Russia; For example, over the past decade, Russia has frequently manipulated social media in relation to Ukraine,” said political economist Alexandra Cirone, an assistant professor of government at the College of Arts and Sciences (A&S), who teaches a class on fake news and disinformation. “But an autocratic country trying to do it to a democratic country in the middle of an election – we are breaking new ground. It is surprising how well these strategies can be transplanted.”
Cirone and Will Hobbs, assistant professor of psychology at the College of Human Ecology and Government at A&S, are co-authors of the article “Asymmetric Flooding as a Tool for Foreign Influence on Social Media,” published March 25 in Political Science and Research Methods became .
Previous research has shown that the Kremlin-affiliated Internet Research Agency (IRA) used thousands of troll accounts on Twitter and other social media platforms to overwhelmingly endorse former President Donald Trump’s campaign against Hillary Clinton support in an attempt to deepen the social divisions between conservatives and liberals. This work has focused on partisan messages.
But in 2018, Twitter released a key dataset with more than 10 million tweets sent from 3,841 IRA accounts. (These accounts represent human-controlled Russian operators, or “trolls,” as opposed to computer-controlled accounts, or “bots.”) The dataset contained apolitical content that was not scrutinized as extensively as the partisan messages.
The researchers asked what was the non-political content doing there?
To analyze the data, the authors used a new text analysis method developed by Hobbs, pivoting text scaling, which can categorize short tweets and the topics of that content over time. He also used his expertise in digital politics and autocratic regimes to spot a pattern he and others have observed in autocratic regimes in the past: the flooding of users’ feeds with non-political content, Hobbs said.
“In China, pro-government users could flood social media feeds with Chinese history or inspirational quotes,” Hobbs said. “So what could Russian trolls in the US use? Entertainment content emerged from the automated analysis method.”
Keywords in these tweets included “hip hop”, “remix”, “rapstationradio”, “nowplaying” and “indieradioplay”.
By creating a timeline of IRA embassies, they found that left-leaning IRA trolls were posting large amounts of entertainment content in their artificial liberal community, turning away from political content late in the election campaign. At the same time, conservative trolls targeted their community with increasing political content. The attempt would have encouraged the right to vote and the left to ignore politics.
Hobbs notes that the mechanism does not mean that a Twitter user sees a music tweet and automatically stops thinking about politics. When a troll suddenly posts a significant amount of music tweets in a feed, they crowd out other content.
“If someone is posting a lot of entertainment content and your feed is only showing 10 posts at a time, then eight posts that aren’t politics now pushes everything else down in the feed,” Hobbs said.
The results underscore that entertainment content is not as benign as we might think. “They have a lot of strategies to fall back on, and entertainment content can be useful for a few goals,” Hobbs said.
Cirone added, “You might think you’re clicking on a cat meme, but in reality you might be putting a troll network in your feed that might later start posting divisive content or monetizing their followers.”
More important are the implications for the upcoming midterm elections, they said.
“There will be all sorts of bad actors trying to meddle at halftime, especially the competitive races. All of them could use entertainment content to distract or impersonate regular users,” Cirone said. “The fact that the average person wouldn’t associate music with Russian trolls means they’re doing a good job.”