F-Secure’s analysis of 24 million tweets exposes inorganic activity in Twitter’s Brexit conversations

Facebook, Twitter, and other social media platforms play a big role in shaping popular opinion. But they’ve opened a pandora’s box of potential disinformation and manipulation of the public. After investigating 24 million tweets related to Brexit, cybersecurity provider F-Secure has identified efforts to amplify pro-leave Brexit views by far-right Twitter users based outside of the United Kingdom.

The research found examples of inorganic activity on both sides of the Brexit debate (the “leave” and “remain” communities). But this activity was far more prominent in “leave” conversations. According to the research, the goal of these efforts was to amplify right-wing populist views around Brexit and other political issues, such as the Yellow Vest protests in France.

“The activity we found happening on the ‘leave’ side of the Brexit conversation was quite different from the more organic appearance seen in the ‘remain’ conversation. And inorganic activity, in relation to political movements and events, can sometimes be indicative of astroturfing or the spread of disinformation,” said Andy Patel, a senior researcher with F-Secure’s Artificial Intelligence Center of Excellence. “At the very least, our research shows there’s a global effort amongst the far-right to amplify the ‘leave’ side of the debate.”

Evidence for the conclusions cited in the research includes:

1. The top two influencers in the pro-leave community received a disproportionate number of retweets, as compared to influencer patterns seen in the pro-remain community

2.  The pro-leave group relied on support from a handful of non-authoritative news sources

3.  A significant number of non-UK accounts were involved in pro-leave conversations and retweet activity

4.  Some pro-leave accounts tweeted a mixture of Brexit and non-Brexit issues (specifically #giletsjaunes, and #MAGA)

5.  Some pro-leave accounts participated in the agitation of French political issues (#franceprotests)

The data consisted of 24 million tweets from 1.65 million accounts collected between December 4, 2018, and February 13, 2019 that contained the word “brexit.” The research analyzed the data looking for signs of suspicious activity, including evidence of botnets, disinformation campaigns, astroturfing, or another inorganic phenomenon.  

While the research included a substantial amount of data regarding Brexit, Patel says it’s only a starting point for this type of investigation, and hopes others continue to build on the findings.

“Social networks generate huge volumes of data. Finding noteworthy trends and phenomena in this data can be complicated, resource-intensive work. But social media is an important source of news and information that many people are still learning how to use. We really hope people see this research, see how much more there is to learn, and start building on our methods to create new, better understandings of what happens on social media.”