We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 - 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!
As if Twitter’s reputation hasn’t been battered enough, a new study sheds light on how the social media platform can be hijacked by bots to spread political disinformation during election campaigns.
A researcher at the University of Southern California found that almost 20 percent of Twitter bots that were engaged in spreading propaganda against Emmanuel Macron during the recent French presidential election had been used to spread misinformation in favor of Donald Trump last year during the U.S. elections.
Many of those bots stopped tweeting after Trump won last November and only began tweeting again in the weeks ahead of the French elections to promote the “Macron Leaks,” an unverified hack of his campaign’s emails.
Bottom line, according to the study by Dr. Emilio Ferrara, a research assistant professor at USC Computer Science Department: “Account usage patterns suggest the possible existence of a black-market for reusable political disinformation bots.”
In the case of the French elections, the bots were not effective because they were primarily reaching an audience outside of France, Ferrara writes.
Still, the notion that money can buy a bot army to sway public opinion for or against a candidate is another blow for a social media platform that once harbored pretensions of being a democratizing force for political speech.
Ferrara’s study is called: “Disinformation and Social Bot Operations in the Run Up to the 2017 French Presidential Election.” The study was published as an open source document, and the full version is here.
To conduct the study, Ferrara collected 17 million tweets related to the French political campaign. Using “machine learning techniques and cognitive behavioral modeling,” he then separated human accounts from those of bots:
…some of the bot accounts we uncovered were created at the beginning of November 2016, shortly before the 2016 U.S. Presidential election, and used only for a week to support of alt-right narratives; then they ‘went dark’, showing no activity till early May, in support of alt-right agenda and the MacronLeaks disinformation campaign in the context of the 2017 French Presidential election.
Those zombie Twitter accounts were tweeting stuff like links to alt-right news media organizations that were hyping the Macron leaks, links to an online archive of the documents, and URLs pointing to well-known far-right fake news websites.
These accounts can rapidly gain followers through retweeting, favoriting tweets, and automatic follow and follow-back mechanisms.
If you’re looking for a hopeful sign, the study notes that out of the 15 most active social bots tweeting anti-Macron messages, Twitter deleted 4, suspended 7, and quarantined 2. So it seems the company is taking some action, though it couldn’t do much in advance of the elections.
Still, the overall prognosis offered by Ferrara about the role of social media is not particularly cheerful.
“The adoption of automated devices such as social bots in the context of disinformation campaigns is particularly concerning because there is the potential to reach a critical mass large enough to dominate the public discourse and alter public opinion,” he writes. “This could steer the public’s attention away from facts and redirecting it toward manufactured, planted information.”
VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn more about membership.