It was the biggest political upset in American history. And the shockwaves of Donald Trump’s unexpected election victory are now being felt, with the slew of executive orders culminating in the 120 day ‘Muslim ban’ which led to chaos at airports over the weekend.
While this shocking act is, understandably, dominating the news agenda, the subject of exactly how he pulled off his unlikely win is still a hugely fascinating topic – and a stunning new article on Motherboard has begun to go viral on the internet which purports to describe one of – if not the – key mechanics of his victory.
German investigative journalists Hannes Grassegger and Mikael Krogerus looked at the role ‘Big Data’ had to play in Trump’s campaign. Big Data is the term used to describe the way that we digitally document virtually every part of our lives and, whether willingly, or inadvertently, make that data available for others to study, use and – potentially – misuse.
Specifically, they have looked at the role that mysterious UK data analysis firm Cambridge Analytica played in the Trump campaign. Led by CEO Alexander Nix, the company – which was also involved in the UK referendum campaign on the Leave side – using a vast array of data and sophisticated modelling, seemingly converts Facebook likes, together with credit reference information, supermarket club card data and land registry information, to create functional demographic data which can then be used to target highly-relevant, specific adverts.
In short: the data which you’ve given out willingly, added together with data freely available to buy or access freely, means that a computer can predict exactly what sort of person you are, and what specific issue is likely to drive you to vote, or not vote.
The methodology has been known about for a while – in January 2015 Wired wrote a feature entitled ‘How Facebook knows you better than your friends do’. It explains that, the genesis of the research began with a 100 item personality questionnaire that went viral on Facebook back in 2007. The questions answered were designed to identify five key personality traits (the ‘Big Five’, with the acronym OCEAN):
Openness (how open you are to new experiences?)
Conscientiousness (how much of a perfectionist are you?)
Extroversion (how sociable are you?)
Agreeableness (how considerate and cooperative you are?)
Neuroticism (are you easily upset?)
It then cross-referenced the scores people achieved with the Facebook ‘likes’ that they shared with at least 20 other respondents. Thus, ‘likes’ could be connected to certain personality traits. The virality of the test – which people willingly did, and before the days that this information might be kept private – meant that the researchers, Michal Kosinski and David Stillwell at the Psychometrics Centre at Cambridge University, had access to information from millions of people. In Motherboard’s words: “the largest dataset combining psychometric scores with Facebook profiles ever to be collected”.
In 2012, Kosinski proved that on the basis of an average of 68 Facebook “likes” by a user, it was possible to predict their skin color (with 95 percent accuracy), their sexual orientation (88 percent accuracy), and their affiliation to the Democratic or Republican party (85 percent) – and even, as a 2013 Wired article showed, how intelligent you are based on your liking of curly fries. But it didn’t stop there. Intelligence, religious affiliation, as well as alcohol, cigarette and drug use, could all be determined. From the data it was even possible to deduce whether deduce whether someone's parents were divorced. With more likes comes more accuracy – to the point that with 300 likes, the system would know a person better than their spouse.
As Motherboard noted, “what Kosinski had invented was sort of a people search engine”, and the scene was set for someone to exploit this.
What appears to have occurred is that Aleksandr Kogan, a young assistant professor in the psychology department, had informed a mysterious company with a labyrinthine ownership structure named Cambridge Analytica, of Kosinski’s methods and then reproduced them. First, they used them on the Leave campaign, then it worked with Ted Cruz, turning a relative unknown into one of the frontrunners for the Republican candidacy before, finally, the company was hired by the Trump team in June 2016.
The detail can be found in the Motherboard article, but essentially, Trump’s team used this to double down on incisive, specific, targeted Facebook adverts, both to energise Trump supporters but also, crucially, to depress the Democrat vote, with Trump’s contradictory, confused mixed messaging paradoxically becoming a useful tool – adverts could simply pick whichever comment Trump had made which would appeal to a particular personality type, and ram it home. Motherboard makes the astonishing claim that “Trump’s team tested 175,000 different ad variations for his arguments, in order to find the right versions above all via Facebook”.
As Cambridge Analytica CEO Alexander Nix says in a September 2016 presentation, "pretty much every message that Trump put out was data-driven. We can address villages or apartment blocks in a targeted way. Even individuals." He also claims that “We have profiled the personality of every adult in the United States of America – 220 million people,”
Additionally – and particularly terrifyingly – the data could be used the other way round: canvassers for Trump could know exactly what type of the personality the occupant of a house was before they even knocked.
Trump spent, relatively, far more on digital campaigning compared to TV than his rival Clinton, and far less overall.
Illegal? No. Clever? Certainly. Rather terrifying? Without doubt. For one side to have access to this huge capability and not the other opens up the possibility of a completely unequal balance of campaigning power.
As Kosinski hears accusations that his work has led to the election of a tyrant into the White House, he says simply: “This is not my fault. I did not build the bomb. I only showed that it exists.”
In 2013, he was quoted in Wired as saying, “I am a great fan and active user of new amazing technologies, including Facebook. I appreciate automated book recommendations, or Facebook selecting the most relevant stories for my newsfeed. However, I can imagine situations in which the same data and technology is used to predict political views or sexual orientation, posing threats to freedom or even life."
Has his prediction come to pass?