People want to know whether this seismic political shift and its profound effects were somehow the result of manipulation or outside meddling.
That people are attempting to manipulate social media to advance their own political agendas is unquestionable. A recent HOPE not hate report, Bots, Fake News and the Anti-Muslim Message on Social Media, provides concrete evidence that so-called ‘bot armies’ are being used to amplify anti-Muslim messages on Twitter.
A report published by Marco Bastos and Dan Mercea of City University, titled The Brexit Botnet and User-Generated Hyperpartisan News, analysed 10m tweets between 10 June and 10 July 2016 that referenced the referendum using relevant hashtags.
They found that an army of 13,493 fake Twitter accounts posted almost 65,000 times about the referendum, only to vanish soon after the vote, while an additional 26,538 suddenly changed their name. Bastos said that they “believe these accounts formed a network of zombie agents”.
Interestingly bots were eight times more likely to tweet Leave slogans than other Twitter users and out of 794,949 users, only 37% (30,122) were located in the UK, raising the question of possible foreign involvement in manipulation.
Russian Influence?
Unsurprisingly, the question of possible Russian interference in the referendum is a deeply complex one, shrouded in rumours and claims to evidence that are hard to substantiate.
For example, Aaron Banks, founder of the unofficial Brexit campaign Leave.EU, bragged in his book The Bad Boys of Brexit, how he and the campaign’s Director of Communications Andy Wigmore had a meeting with the Russian First Secretary at the Embassy (“in other words, the KGB’s man in London”) at the ambassador’s private residence in Kensington Palace Gardens. The truth and relevance of this to the question of Russian interference, however, remains underdetermined.
In terms of more tangible evidence, what we can say is that while there was indeed Brexit-related bot activity on social media during (and after) the debate, there is little evidence that it was predominantly directed by Moscow, or importantly that it had any noticeable effect.
Researchers at Swansea University and the University of California, Berkeley collected just over 28 million pro-Remain and pro-Leave tweets between 24 May and 17 August 2017 and found that 20% of the accounts were bots.
They tracked 156,252 Russian accounts that mentioned #Brexit and found they posted almost 45,000 messages pertaining to the EU referendum in the 48 hours around the vote.
They also found that “During the Referendum day, there is a sign that bots attempted to spread more leave messages with positive sentiment as the number of leave tweets with positive sentiment increased dramatically on that day.”
However, when attempting to gauge the impact of these Russian-linked tweets, it is important to note that the vast majority (39,000) were tweeted the day after the vote, thereby nullifying their actual impact on the referendum.
Other researchers at the Neuropolitics Research Lab at the University of Edinburgh searched through over 62 million tweets, cross-referencing them with a list of 2,752 Russian troll accounts produced by Twitter identified as being active in the US 2016 election and found 3,468 tweets from 419 users on the list. Just 400 of these tweets, published by 38 users, were on the date of the EU referendum while a further 432 tweets were in the week of the referendum and made by 58 users. Once again, however, 78% of the tweets came after the date of the referendum.
Worryingly though, the Edinburgh-based researchers found that one of the 419 accounts identified as operating from the now infamous St Petersburg-based ‘troll farm’, the Russian Internet Research Agency, later attempted to spread anti-Muslim sentiment during the Westminster Bridge terror attack.
In December 2017, the Oxford Internet Institute also published a relevant report, Russian Involvement and Junk News during Brexit, that found 105 Russia-linked Twitter accounts publishing Brexit related content in the run-up to the referendum. They tweeted 16,000 times in two separate weeks.
However, researchers concluded that “Russian Twitter accounts shared to the public, contributed relatively little to the overall Brexit conversation”, adding that “Russian news content was not widely shared among Twitter users” and that “only a tiny portion of the [tweeted] YouTube content was of a clear Russian origin”.
When it came to social media advertising spending linked to Russia during the referendum, evidence is even less conclusive. A Facebook investigation into Russian influence found just $0.97 of ad spending by Russian based “troll army” meaning just 200 people in Britain saw the adverts. According to Facebook, “This amount resulted in three advertisements (each of which were also targeted to US audiences and concerned immigration, not the EU referendum) delivering approximately 200 impressions to UK viewers over four days in May 2016.”
Meanwhile, Google found no evidence of paid activity linked to the referendum, while Twitter only found six adverts by the Russian state broadcaster Russia Today. Interestingly, the Oxford Internet Institute found that just 0.6% of the links shared in tweets with Brexit hashtags were from Russian sources like the state controlled Russia Today or Sputnik.
It is perhaps unsurprising then that when asked about Russian interference, the ex-head of MI6, Sir Richard Dearlove, said, “I’ve not seen anything that convinces me at all that the Russians intervened significantly in the Brexit referendum.”
We have to be led by the evidence and while, in the case of Brexit, it is clear that large numbers of bots were active on social media during the referendum, many of which spread highly polarizing political content, there is little solid evidence they had much effect and even less to indicate significant Russian involvement.
No Excuse
None of this is to say that we should not be vigilant regarding the impact of possible Russian influence and social media manipulation. These are vital issues that demand to be taken seriously, and will only become more important as social media becomes ever-more ubiquitous in people’s lives. Such questions pose fundamental questions about the very health of our democracies.
Yet, we must also be careful not to take proof of attempted manipulation as proof of its impact. Much more research is required to determine the effect of such social media manipulation on the political process, even where its existence has been established.
A recent collaborative report by researchers at Princeton University, Dartmouth College and the University of Exeter titled the Selective Exposure to Misinformation, explored the impact of fake news during the 2016 US elections. While accepting that “selective exposure to politically congenial content [is] likely to extend to misinformation”, the report found, in the words of the New York Times, “Wide Reach but Little Impact”. While one in four Americans saw at least one false story, it was found that people still consumed far more real news from legitimate sources. False stories made up just 1% of Clinton supporters’ news intake and just 6% of Trump supporters’.
There is a danger that legitimate concerns about fake news, Russian interference and social media manipulation could become monocausal excuses for the problems we face. Even in the US, where the evidence is much clearer, over-focusing on these issues risks conveniently exonerating progressive movements of our failures, while also ignoring the role of societal and structural racism and the central role of domestic actors that consciously push the politics of hatred and division.
A dismal turnout is one of several concerns facing leader Paul Golding after another poor national event, as we identify a puzzling discrepancy around the…
