Cybercrime

Research Shows Twitter Manipulation in Weeks Before EU Elections

This is an age of large scale political social engineering through social media, both by advertising and the presentation of misleading data. International social engineering became frontpage news with the 2016 US presidential elections, but has not abated since. Researchers with the Sherpa project analyzed the use of social media as a recommendation system — specifically Twitter — ahead of the European elections in May 2019.

<p><span><span>This is an age of large scale political social engineering through social media, both by advertising and the presentation of misleading data. International social engineering became frontpage news with the 2016 US presidential elections, but has not abated since. Researchers with the Sherpa project analyzed the use of social media as a recommendation system -- specifically Twitter -- ahead of the European elections in May 2019.</span></span></p>

This is an age of large scale political social engineering through social media, both by advertising and the presentation of misleading data. International social engineering became frontpage news with the 2016 US presidential elections, but has not abated since. Researchers with the Sherpa project analyzed the use of social media as a recommendation system — specifically Twitter — ahead of the European elections in May 2019.

The Sherpa project comprises a consortium of 11 partners (academia, industry, civil society, standards bodies, ethics committees) across six European countries. Its purpose is to analyze and understand the ways in which smart information systems and big data can impact ethics and human rights issues. As part of this project, F-Secure’s Andy Patel — a researcher in its artificial intelligence center of excellence — analyzed political tweets in the weeks leading to the EU elections.

After examining tweets that included official EU election hashtags, such as ‘#EUElections2019’, “We stumbled upon some highly political tweets that received a great deal of suspiciously inorganic engagement,” reports Patel. Inorganic engagement can be defined as manipulated engagement, while organic engagement comes from the normal interaction of friends and interested parties. The former is typified by a high percentage of retweets — often from fake accounts — while the latter tends to have a higher percentage of ‘likes’.

The report focuses on two particular inorganic tweets. The first, posted by NewsCompact, shows a video of “A muslim #migrant destroying a statue in #Italy, because part of the body is showing.” In fact, the video was taken in Algeria; but nevertheless, it was retweeted by thousands of far-right accounts.

The second tweet came from an account operated by PartisanDE: “More evidence how inhuman #Macron’s #France is in 2019. A disabled protester was humiliated by #regime forces, her walker was knocked out of hand. She fell down.” This tweet received more than 6,000 retweets. 

These two accounts are connected — the operator of PartisanDE identifies as the CEO of NewsCompact. Both have a high number of following accounts that were created within the last 18 months. “They share a great deal of politically inflammatory content, racist content, and mistruths,” writes Patel. “The accounts also share many URLs to sources such as VoiceOfEurope, RT, Sputnik, and some other non-authoritative ‘news’ domains.”

The research discovered 200 accounts engaged in posting political content relating to the EU, all of which had been created on 20 and 21 May 2019 (two days before the UK started the EU voting). All these accounts retweeted a PartisanDE tweet saying “#Brussels poster boy #Macron now has a disapproval rating of a staggering 68% after weeks of chaos following his dreadful mishandling of the #GiletJaunes.” This tweet was stamped with the leave.eu logo — the official campaign operation for the ‘leave’ side during the Brexit 2016 referendum.

At the time of compiling the report, this tweet had received 388 retweets, including the 200 identified new accounts. Looking at those 200 accounts, Patel discovered that they have also retweeted legitimate “verified accounts (such as justinsuntron, paulwrblanchard, and Mark_beech), selling products (including books an health supplements), people in the music industry, a fashion blogger, finance ‘experts’, ‘social influencers’ (surprise, surprise), and even ‘CEOs’.”

Advertisement. Scroll to continue reading.

Patel’s conclusion is that not only political factions, but apparently verified accounts can and do freely purchase retweets, followers, and therefore influence. Nevertheless, they also showed a strong political element that seemed to be aimed at galvanizing the right wing throughout Europe in the days prior to the EU elections. These two accounts (PartisanDE and NewsCompact), writes Patel, “are heavily fabricating engagement and, at least PartisanDE is also purchasing retweets.”

Interestingly, the ‘muslim migrant’ tweet and the ‘disabled protester’ tweet seem to be specifically targeting Italy and France respectively. In both countries, the far-right parties won the biggest slice of votes in the subsequent election — Matteo Salvini’s League party in Italy, and Marine Le Pen’s National Rally outscoring President Macron’s party by 23.31 percent to 22.41 percent.

In the UK, Britain’s Brexit party led by Nigel Farage, formed only a few months ago, was the clear winner with 28 seats. While this current research from Patel does not highlight UK political tweets, Patel conducted similar research examining Brexit in March 2019. The basic conclusion then was that loosely related far-right groups from both within and outside Europe were still attempting to keep the Brexit flag flying.

The research does not attempt to find the ultimate source of this political tweeting, whether influenced by foreign governments or just right-wing agitators. Nor can it say whether such tweets have any effect on votes. All that can be said is that the use of social media by nearly all political parties for large-scale political social engineering is continuing, and shows no sign of abating despite public avowals from the social networks that they are clamping down.

Related: Facebook Closes Hundreds of ‘Inauthentic’ Russia-Linked Pages 

Related: Facebook Uncovers Political Influence Campaign Ahead of Midterms 

Related: Google Blocks Accounts in ‘Influence Operation’ Linked to Iran 

Related: How China Exploits Social Media to Influence American Public 

Related Content

Copyright © 2024 SecurityWeek ®, a Wired Business Media Publication. All Rights Reserved.

Exit mobile version