Part One of this series discussed the work of a researcher at the University of Cambridge Psychometrics Centre named Michal Kosinski. He developed a computer model which could predict key elements of a person’s personality using Facebook data as inputs. Aleksandr Kogan reached out to Kosinski on behalf of a mysterious company called Strategic Communication Laboratories (SCL), and was interested in licensing the model. Kosinski was concerned that his research may end up in the wrong hands, and refused the offer.
However, it did not take long for SCL to develop its own computer model based on Kosinski’s research.
According to documents obtained during an investigation by The Guardian, in early 2014, mere months after Kosinski declined SCL’s offer, they decided to partner with Kogan instead. Kogan learned of Kosinski’s method, and began to build his own database by paying Amazon Mechanical Turk (MTurk) workers about US$1 each to take the OCEAN test and share their Facebook data. They were told that this data was being used “only for research purposes” and would remain “anonymous and safe,” but they were not told that Kogan was also collecting the same data from their pool of friends.
Kogan was reported by multiple participants for violating Amazon’s terms of service, and the program was later shut down. However, by that time Kogan’s dataset had already ballooned significantly in size. He set out to construct his own computer model, and like Kosinski began correlating Facebook data with responses to the OCEAN test. That summer, Kogan’s company, Global Science Research (GSR), boasted on LinkedIn that it possessed a dataset composed of more than 40 million individuals across the United States. Moreover, GSR claimed to have generated detailed characteristic and personality trait profiles for each of these individuals.
When Kosinski discovered that Kogan had secretly registered a company that was doing business with SCL, he immediately suspected that Kogan either copied or rebuilt his model, and then sold it to the election management agency. He abruptly broke off contact with Kogan and reported him to the director of the University of Cambridge Psychometrics Centre. A scandal erupted within the university, and Kogan moved to Singapore and assumed the name Dr. Spectre. Documents obtained by Das Magazin reveal that SCL did indeed learn intimate details of Kosinski’s methods from Kogan.
Once the data was in their hands, SCL formed a company called Cambridge Analytica, which built their own version of Kosinski’s computer model. They then sought out wealthy conservative investors such as Robert Mercer, who pledged over US$10 million.
The model itself was in its infancy, but it showed great promise. However, it needed to be refined and it was hungry for a larger dataset, thus prompting Cambridge Analytica to go on a “data shopping spree.” They went looking on the open market, scooping up data about shopping habits, land ownership, where people go to church, which magazines they subscribe to, and how they spend their time online. All of this data, and more, is for sale from data brokers and third-party organizations that sell information about their users. Cambridge Analytica combined this data with electoral rolls (lists of registered voters) and publicly available online data such as Facebook likes, then fed them into its predictive personality model.
What did they intend to do with this model? Why, win an election of course, but which one? In November 2015, Nigel Farage announced that “Leave.EU,” the more radical of the two Brexit campaigns, had enlisted the aid of Cambridge Analytica to manage their online campaign. They used something called “behavioural micro-targeting” to deliver political messages tailored to appeal to a specific person’s personality profile.
However, in December 2016, Cambridge Analytica sent a letter to the Observer denying that it had a hand in the Leave.EU campaign. They stated that Cambridge Analytica “is a US company based in the US. It hasn’t worked in British politics.” Interestingly, Cambridge Analytica’s website claims that it is a British company based in London.
Moreover, in February 2017 The Guardian interviewed Andy Wigmore, communications director for the Leave.EU campaign. He confirmed that Cambridge Analytica had indeed worked with their campaign, and taught them how to collect massive amounts of data from Facebook profiles, build a psychometric profile, and then send them targeted political advertisements.
However, he said, they did not “employ” Cambridge Analytica. “No money changed hands… They were happy to help.” And Wigmore explained why: “Because Nigel [Farage] is a good friend of the Mercers. And Robert Mercer introduced [Cambridge Analytica] to us. He said, ‘Here’s this company we think may be useful to you.’”
In the interview, Wigmore also explained how a Facebook “like” was their most “potent weapon”. “Because using artificial intelligence… tells you all sorts of things about that individual and how to convince them with what sort of advert… And then you follow them. The computer never stops learning and it never stops monitoring.”
Cambridge Analytica’s wealthy conservative investors were pleased with the Brexit results, but this was just a trial run for their Big Data approach to political marketing. They had much bigger fish to fry: they had their eyes set on winning the upcoming US election.
This is Part Two of a four-part series detailing the rise of a new type of propaganda machine; one tailored to fit a modern, inter-connected, and technologically dependent society. Part Three will delve into Cambridge Analytica’s role in the 2016 US election, the methods they employed, and their power structure.
Photo: Professional Businessmen, by Unknown via Source. Public Domain.
Disclaimer: Any views or opinions expressed in articles are solely those of the authors and do not necessarily represent the views of the NATO Association of Canada.