We’ve been able to see just a handful of the ads that were blanketed across social media during and after the referendum campaign. Vote Leave had found their niche and they were hitting voters hard with the same lines over and over, digitally and on TV and radio. Cummings had realized there was no oversight on these adverts and was not going to let truth stop him from causing the biggest upset in British politics. One advert seen between 2 and 4.9 million times was targeted almost exclusively (99 per cent) at English Facebook users — and included the claim that: ‘EU protectionism has prevented our generation from benefiting from key global trade deals. It is time we unite to give our country the freedom to be a prosperous and competitive nation!’ BeLeave were slightly more subtle with their approach, urging target recipients to back a ‘fair immigration system’ or an ‘Australian-style points-based system’ but without making any direct references to any specific non-EU countries. BeLeave also created a handful of ads which focused on EU regulation of modern technology as a reason to back Brexit. They criticized EU regulation of digital streaming services and ride-sharing apps like Uber, ‘We can’t let EU regulators keep us in the past. Learn More’. Or ‘The EU should not be forcing quotas on streaming. Learn More’.
The most problematic part of these dark ads is the lack of oversight, there is no factual test, regulatory barriers, no ombudsman to complain to if adverts are misleading, disingenuous, or just straight-up lies. That is, unfortunately, a huge part of the internet. We have to use our own judgement to determine the validity of claims we see – ‘Will this supplement in my Facebook feed really make me burn twice as many calories?’ ‘Does this Nigerian Prince really need my help?’ ‘Do HMRC really owe me £13,000?’ Twitter has now banned all political advertising, but it remains to be seen as to whether Facebook will follow suit.
All of these seem easy to decipher, they are obviously extreme examples. But do remember that if no one clicked on these adverts, they would not still exist. Because we are permanently plugged into the news cycle through social media, we have to be on constant alert. It’s exhausting, never-ending scepticism, but it is what is required from all of us in twenty-first century society to remain informed and aware of what exactly is going on in the country. For far too long it has been easy not to care, to watch from afar and assume everything was running smoothly enough. We have out-lived that luxury and to ensure the survival of our very democracy we will be forced to tune in and educate ourselves.
Facebook has since issued a statement declaring that all ads running on its platform display an associated Page — in the top left corner, where users can click through to find out more. Although the relative sizes of the ads when they appear in the news feed means it is possible people could have clicked on an imprint-less ad without noticing it was being run by a Brexit campaign.
Initially, Vote Leave tried to stay clear of an immigration driven campaign because they bought into the idea that Farage’s perceived racism turned off more voters than it brought in. They had to make their campaign more subtle. Initially about taking back control of laws and borders and the cost of immigration, then slowly moving towards a more immigration-focused campaign strategy. You may be tempted to say that this was data-driven, that testing had revealed immigration-based arguments played best. Dominic Cummings and Paul Stephenson discussed the transition to immigration as far back at 2015, they believed that to go hard on immigration too soon would turn off voters at the beginning: ‘It’s there, it’s a massive issue, we can go hard on it at the end – and we probably will go hard on it at the end – but we lose a whole bunch of people if we do it straight away.’41 Farage told journalist Tim Shipman that he felt the moment the campaign changed was when Vote Leave began discussing the Australian points-based immigration system that he had long advocated for, ‘within 48 hours I thought “we’re going to win this”.’42
These dark ads were all able to fall on the fertile ground due to the cultural and political climate we inhabit. The loss of trust in mainstream media led people to look elsewhere for news whilst the algorithms on Facebook, Twitter, Google, and YouTube helped us construct echo chambers and epistemic bubbles that reinforced, radicalized, and fed our own belief systems. The ability for political campaigns to harvest data and target adverts with no oversight allowed social media to be flooded with hidden propaganda. It took advantage of the poor quality of online debate, our inability to objectively assess sources, and, if you were deemed by algorithms from Cambridge Analytica or Facebook to be a potential swing voter, you were bombarded with the same messaging over and over. You might hear ideas repeated in the bar, on your timeline, or discussed on the evening news.
As ideas slid seamlessly from memes, ‘junk news’, or dark ads, to comment sections, through blogs and into the mouths of politicians, they became part of the national political conversation. The source of the original point or comment becomes irrelevant once the discussion has begun. Take the debate over Turkey as an example; the idea was propagated by dark ads on Facebook, repeated by members of the Vote Leave and Leave.EU campaigns and discussed in comment sections, Facebook groups, and links posted all across social media. It was plastered across the front pages and came up time and again in debates on TV and on YouTube. Turkey has been applying to the EU since 1987, the UK has a veto. This is all prefaced on Turkey overcoming human rights concerns, corruption within government and the courts. By 2015, of the 33 chapters of membership, Turkey had only negotiated one successfully – science and research, in 2006.43 Yet, none of this seemed to matter (or even be mentioned). Turkey continued to rattle around as part of the debate. Perhaps most ironically, Brexiteers Boris Johnson and Daniel Hannan were founders of Conservative Friends of Turkey and Johnson declared a few months after the referendum that he would campaign to help Turkey join the EU.
When I first began writing and researching for this book, I had mystical ideas about the power of big data analytics. The data alchemists could descend from their Mount Olympus bestowing superhuman predictive powers upon those with a big enough cheque book. As I read further into the capabilities, I felt slightly underwhelmed at the predictive powers that they possessed. Whilst they can successfully identify your political beliefs, sexual orientation, and hobbies, as well as likes or dislikes, their ability to predict personality OCEAN traits was barely above random, hovering around 60 per cent. For a brief moment, my illusions had been shattered. Perhaps all the claims about the power of big data to influence politics were lies. Perhaps Cambridge Analytica was selling snake-oil.
In Outnumbered, David Sumpter expresses serious doubts over the actual impact of microtargeting on social media. He reasons that this is no more powerful than human judgement44. Cambridge Analytica achieved two upsets in 1 year, there is something more than just luck going on here. Its true power lies in its sheer scale and the unaccountability that it provides to political campaigns. They can throw as much money at these adverts as they please with little to no oversight of the content or the source. They can happily flood social media with misinformation on misleading claims. Winning an election or referendum, especially one as partisan and binary as the Brexit vote (or the 2016 presidential election), is achieved by altering the minds of millions of people. The majority of people were already sold on one side or the other, long before the votes were cast. In the case of the Brexit vote, some Eurosceptics had been waiting decades to stick it to the bureaucrats in Brussels. Vote Leave only needed the support of enough of the swing voters, the independents, the undecided. If you invade social media with broadly targeted messaging, inducing fear and resentment amongst those who you believe to be receptive to your ideas, success need not be measured by turning millions of voters with a 100 per cent success rate, it simply needs a few thousand here and there. Bombard the population with your message over and over and you’re likely to win over some converts, even if it is only temporarily, i.e. until the vote takes place. Anti-EU sentiment in the UK peaked during the week of the referendum vote when Vote Leave was spending most of their money. With the ability to reach into the Facebook feed of almost every eligible voter in Britain, it is not accuracy but scale that becomes important. Something around 1 billion impressions might be enough.
It would be lovely to be able to breathe a sigh of relief and tell yourself that the dark Brexit ads were confined to during the referendum. Sadly not. In the year to October 2018, a website known as mainstreamnetwork.co.uk spent an estimated £250,000 on anti-Chequers (the rough agreement put together by Tories during a marathon cabinet retreat in 2018) pro-Brexit adverts. One unknown website essentially footed half the bill that Facebook was sanctioned with by the ICO for data breaches and misuse during the Brexit referendum. It was insignificant for Facebook, they earn billions a year selling you and your data, so why on earth would they stop now? There are no legal requirements to make their advertising process more transparent and seemingly none on the way with the British government all but paralyzed dealing with Brexit itself. This is still happening and will continue to happen until we legislate against it.