This Thursday, the UK finds itself in the odd position of voting for politicians in an organisation it won’t be part of for much longer, and with the Cambridge Analytica scandal having played a role in the 2016 EU referendum and US Presidential Elections in the same year, some might be wondering whether we are again being coerced into voting a certain way by dubious social media accounts.
However, an Oxford Internet Institute study published today suggests that ‘junk news’ levels have been relatively low in the run-up to the EU Parliamentary Elections. This is particularly the case on Twitter, where under 4% of the stories shared on the platform as the elections approach have been classed as junk news. Researchers say that users of the microblogging site are favouring professional news to fake and misleading stories, although Polish-speaking Twitter users are still circulating 21% junk.
On Facebook, the study notes that users are also preferring mainstream content, but that junk stories tend to be heavily interacted with when they crop up. In some cases, such posts receive four times the interaction of professional content, with unreliable stories about immigration and anti-Muslim themes tending to draw in likes, comments and shares.
Perhaps surprisingly, the study also observes that the role of Russia in spreading fake news on social media is minimal, with more of it originating in the country at which it is targeted.
BBC research in collaboration with the University of Birmingham, however, shows that bot creation is still commonplace. A major spike in new Twitter accounts was observed on Saturday, May 11, with the newly set up accounts tweeting heavily on Brexit and other political themes. Most of them consisted of a name followed by eight numbers, such as @johnie76662158, which tweeted well over 1,000 political posts in 10 days – some of them in support of far-right activist Tommy Robinson – before being shut down by Twitter.
Professor Oleksandr Talavera, who was behind the research, says that the parties creating bots are finding new ways to avoid being detected, so it is never possible to truly say how many accounts are run by bots, but he estimates that one in four or five newly created Twitter accounts with a political focus will be automated.
This all suggests that social media sites are doing a better job of closing the gate to dubious accounts, but the fact that ‘alternative media’ is so often shared suggests that some social media users still lack education on how to tell real news from fake. The next step, especially for Facebook, must be to ensure that users are informed on the reliability of the stories they see, share and interact with, perhaps with the help of fact-checking tools.