Mindfuck Read online

  More to the point, I wanted to help Nix understand that a fancy layer of psychological analysis is a complete waste if all you’re going to do is dump your info out of the sky. I told him that SCL’s projects would be far more effective if they focused on getting more accurate data, building algorithms, targeting specific people based on those algorithms, and using different forms of media than a leaflet or radio. Nix listened pensively, his steepled hands repeatedly tapping his mouth as he thought about what I was saying.

  I also started to realize why the British and American militaries were so bad at winning the proverbial hearts and minds. This cultural and attitudinal information about the population was gathered in silos, often in contractor-led ancillary projects that were not integrated into military strategy until after primary objectives were set—in other words, the culture and experiences of local populations were an afterthought for planners, trumped by personnel and equipment. That needed to change.

  * * *

  —

  AS I PONDERED WHAT DARPA and its British equivalent, the Defence Science and Technology Laboratory (DSTL), were trying to develop with their new social network and digital research programs, my mind wandered to an unexpected but not unfamiliar place for me: fashion. The two fields are not as disparate as it might seem. When a society jerks into extremism, so does its fashion. Think about Maoists, Nazis, Klansmen, and jihadists—what do they all have in common? A look. Extremism starts with how people look and how society feels. Sometimes it creates literal uniforms: Olive tunics and caps with red stars, red armbands, white pointed hoods, polo shirts and tiki torches, MAGA hats. These uniforms, in turn, are incorporated into the wearer’s identity, transforming their thinking from This is what I believe into This is who I am. Extremist movements latch on to aesthetics because so much of extremism is about changing the aesthetics of society. Oftentimes much of what is promised is not about any tangible policies, but rather a new look and feel for a place or culture.

  When I was sixteen, I dyed my hair one day to a mulberry shade of purple. There was no particular reason for choosing the color aside from the fact it was eye-catching and I liked it. It also landed me in the principal’s office, for violating the school dress code. Far from being upset or intimidated, I was totally at ease. Finally, I was talking to the principal about something other than “disability accommodations.” I was told that I needed to change my hair back to a “normal” color. I refused. The principal was not happy, and the tension over my hair persisted until I left school. When I was still using a wheelchair, I spent a lot of time thinking about fit—fitting through doors, fitting in with my peers, finding clothes that fit. Computers were one passion, but fashion became another, for more than one reason. It was partly about feeling included. But it was also about being seen. As I sat at waist level, noticing the buttons, the cuts, the creases, the bulges, and the folds in the clothes of classmates, I felt invisible to them. With purple hair, I was seen. And then, when the principal asked that I go back to “normal” hair, he was ordering me to become invisible again. That was when I understood how powerful—and revealing—a look can be.

  As I worked with the five-factor model of personality for the Lib Dems, I started to think more deeply about personality as a construct. Politics and fashion were built on the same foundation, I realized, in that they were both based on nuanced constructs of how people see themselves in relation to others. Fashion is an ideal window into personality, as choosing what to wear (or not) is a decision we all make on a daily basis. People in all cultures make choices about how to adorn their bodies, from the mundane to the extravagant. We all care about what we wear—even the straight old man from Minnesota who never wears anything but a gray T-shirt and jeans. He doesn’t think he cares about how his clothes look, until you offer him a kimono or a dashiki.

  I distinctly remember the final meeting I had with my university tutor at LSE, when he asked what I planned to do next. No doubt he expected to hear that I was continuing in politics or applying to a fancy corporate law firm. Instead I told him I was going to fashion school. Silence. Eyebrows raised and clearly disappointed, he unconsciously shook his head. Fashion? As in clothes? You really want to study clothes? But to me, fashion and politics are both, at their core, about cycles of culture and identity. To my mind, they’re essentially two manifestations of the same phenomenon—a conviction that would become central to what we created at Cambridge Analytica.

  Fashion has always played a role in my life, and it really was the thing that let me become more comfortable with myself. When I left school and relocated to Montréal, I was moving more and more outside of a wheelchair, but that sense of not being attractive or desirable stayed with me. Wandering around one weekend, I found myself in a vintage book shop and discovered a frayed nine-year-old issue of Dazed & Confused in a stack of old magazines. It was a 1998 issue, with the cover line FASHION ABLE? and showing a model with two prosthetic legs. It was guest-edited by Alexander McQueen, and inside were brilliant pictures of bodies that looked different but beautiful. After looking through that issue, I started experimenting with clothes and going out more. Montréal is the sort of place that will change you if you let it. I found myself drawn to the drag bars and admiring a form of dress that can be glam and sumptuous while mocking and upending conventional notions of beauty, bodies, and gender. Drag inverted my thinking. It showed me how to not just defy these social norms, but to laugh at them and simply be who you want to be on your own terms.

  In my early years in London, many of my friends were fashion students at Central Saint Martins, which is one of the constituent colleges of the University of the Arts London. I started as a student at UAL and ended up working under the supervision of Carolyn Mair, who had a background in cognitive psychology and machine learning. Dr. Mair wasn’t a typical fashion professor, but the match made sense, as I wasn’t a typical fashion student. After I explained to her that I wanted to start researching fashion “models” of another kind—neural networks, computer vision, and autoencoders—she convinced the university’s postgraduate research committee to allow me to commence a Ph.D. in machine learning rather than in design. It was around this time that I also began my new job at SCL Group, so my days fluctuated between fashion models and cyberwarfare. I was keen to dive into my academic research on cultural trends, so I told Nix that I did not want to work for SCL full-time, and that if SCL wanted me, they would have to accept that I would be continuing my Ph.D. in parallel to their projects. Nix agreed to this arrangement and SCL eventually agreed to cover my tuition fees, which felt like a godsend for me, since as an international student I had to pay the highest rate of tuition.

  These two domains serve each other well, as understanding culture can equip you to unpack the dynamics of extremist movements more than purely looking at their professed ideologies can. At SCL we would watch countless numbers of radical jihadist propaganda videos, and we noticed that, beyond the violence of the clips that make it onto the news, there was a rich and well-articulated aesthetic to their style of content. Cool cars were showcased. There would be music. There was a defined masculine look to their idealized heroes, and some of the videos looked almost like clips from reality TV. The irony was that they tried to position their backward ideology as somehow modern or futuristic in a way that echoed the old Italian Futurists’ promotion of a fascism for tomorrow—that it was the most expedient gateway to modernity. These films were propagating a grotesque cult of violence and hate, but beyond that, they also formed part of their culture. Their style was self-indulgent and naïvely romantic, and it bordered on kitsch. Even terrorists have pop culture.

  Around this time, in September 2013, I distinctly remember thinking, How cool is this? I get to work in culture, but not just for someone’s branding campaign. I get to work in culture for the defense of our democracy. The military just used different terms—modeled influence attribution or target profiles observed acting in c
oncert. But in fashion, we just call that a trend. Dressing in concert. Hashtagging in concert. Listening in concert. Going to a concert in concert. The cultural zeitgeist itself is just people acting in concert. And these kinds of trends, I was sure, could be discerned in the data. Through online observation and profiling, we wanted to try to forecast these movements’ adoption life cycles, their early adopters, their diffusion rates, their peaks.

  In my first weeks at SCL, I began looking into how to digitize and transform the traditional tactics of information operations. This was what the firm was most interested in at the time, as it realized that there was a critical capacity gap in many NATO militaries that it could fill (and profit from) if it developed new ways of merging propaganda with ad tech. This involved exploration of what research we could draw upon about mapping out this new digital domain, such as acquiring new information sources from clickstreams and improving the targeting of narratives at their intended recipients through profiling and machine learning. There are obvious inherent complications in weaponizing information. Guns and bombs kill people no matter who or where they are—the properties of physics are global. But an information weapon has to be tailored according to multiple factors: language, culture, location, history, population diversity. If you’re building a non-kinetic weapon designed for scaled perspecticide—the active deconstruction and manipulation of popular perception—you first have to understand on a deep level what motivates people.

  Insurgencies, by nature, are asymmetric, in that a few people can cause large effects. So catalyzing an insurgency within the belligerent’s organization requires first concentrating resources on a few key target groups. This is optimized by good profiling and identifying the types of people who are both susceptible to new ways of thinking and connected enough to inject our counternarrative into their social network.

  The most effective form of perspecticide is one that first mutates the concept of self. In this light, the manipulator attempts to “steal” the concept of self from his target, replacing it with his own. This usually starts with attempting to smother the opponent’s narratives and then dominating the informational environment around the target. Often this involves gradually breaking down what are called psychological resilience factors over several months. Programs are designed to create unrealistic perceptions in the targets that result in confusion and damage self-efficacy. Targets are encouraged to begin catastrophizing about minor or imagined events, and counternarratives attempt to remove meaning, creating an impression of confusing or senseless events. Counternarratives also attempt to foster distrust in order to mitigate communication with others who might hamper the target’s evolution. It is much harder to stay loyal to an existing hierarchy or group when you begin to think that you are being used in some unfair way, or when events seem senseless or purposeless. You become less willing to accept setbacks, take risks, or comply with commands.

  But simply degrading morale is often not enough. The ultimate aim is to trigger negative emotions and thought processes associated with impulsive, erratic, or compulsive behavior. This moves a target from mild or passive resistance (e.g., less productivity, taking fewer risks, rumors, etc.) into a realm of more disruptive behaviors (e.g., arguing, insubordination, mutiny, etc.). This approach has been taken in South America, for example, to provoke disunity among members of narcotics operations, increasing the likelihood of information leaks, defections, or internal conflicts that erode a supply chain. The most susceptible targets are typically the ones who exhibit neurotic or narcissistic traits, as they tend to be less psychologically resilient to stressing narratives. This is because neuroticism can make a person more prone to paranoid ideation, as they tend to experience more anxiety and impulsiveness and place more reliance on intuitive rather than deliberative thinking. People high on the narcissism scale are susceptible because they are more prone to feelings of envy and entitlement, which are strong motivators of rule-breaking and hierarchy-defying behavior. This means these targets will be more likely to develop an exaggerated suspicion of harassment, persecution, victimhood, or unfair treatment. This is the “low-hanging fruit” for initiating the subversion of a larger organization. Later, this learning would serve as one of the foundations for Cambridge Analytica’s work catalyzing an alt-right insurgency in America.

  Let’s be clear: These operations are not some kind of therapeutic counseling; they are a form of psychological attack. It’s important to remember that, in a military context, the target’s personal agency or consent is not a concern. The target is the enemy. The choice for the military is often either to send in a drone to incinerate the enemy or to mess with the enemy’s unit to such an extent that they begin to fight among themselves or get sloppy and make exploitable mistakes. If you are a military commander or an intelligence officer, psychological manipulation is the “light touch” approach.

  With the advent of social media, suddenly military and security agencies had direct access to the minds and lives of guards, clerks, girlfriends, and runners of criminal and terrorist organizations all around the world. What the social data offered was a trail of detailed personal information that previously would have taken months of careful observation to gather. The targets were in effect creating their own dossiers with rich data that could quicken a psychologist’s assessment of their disposition. This spurred a host of research into psychological profiling that could be automated with machine-learning algorithms. These algorithms would allow agencies to widen their net through automation and reach the scale of an old-school leaflet drop, but with the precision of targeted messages. In 2011, DARPA began funding research into psychological profiling of social media users, how anti-government messages spread, and even online deception. Engineers at Facebook, Yahoo, and IBM all participated in DARPA-funded research projects to assess how information is consumed and spread. The Russian and Chinese governments also launched their own social media research programs.

  On my very first day at SCL, Nix asked whether I’d heard of a company called Palantir. He’d learned about it from an unusually well-connected SCL intern named Sophie Schmidt—the daughter of Eric Schmidt, a billionaire and then executive chairman of Google. A few months earlier, as she was finishing up her internship, she’d introduced Alexander to some of the executives at Palantir. Co-founded by Peter Thiel, a well-known venture capitalist in Silicon Valley who was also an independent director of Facebook, Palantir was a massive venture-capital-funded company that undertook information operations for the CIA, the National Security Agency, whose mission is to analyze signals intelligence and data for national security purposes, and the Government Communications Headquarters (GCHQ), the British counterpart to the NSA. Nix was obsessed. He wanted SCL to do what Palantir was doing.

  In my first few months, I worked on small pilot projects in various countries with Brent Clickard, the Cambridge-based psychologist, and a friend of his named Tadas Jucikas. I first met Jucikas at the Royal Automobile Club, a private members’ establishment with bars, squash courts, and billiard rooms, where the upper crust go to socialize and do business. It was founded back in 1897, when driving a car was an exorbitantly expensive hobby, and has retained its air of lavish elegance. When I walked up to the colonnaded clubhouse, I saw Jucikas standing in the lobby beside a bright-red antique racing car, his eyes hidden behind tortoiseshell sunglasses. He was wearing a beautifully cut herringbone jacket with a crisp pocket square. It was all so extra, but I was into it.

  He led me inside the club, where we drank a couple of boulevardiers before moving to the balcony to smoke cigars. Jucikas had grown up in rural Lithuania, where he watched Soviet tanks roll through his town as a boy. He was obviously brilliant. As we sat on the balcony, savoring our cigars and discussing artificial intelligence and data pipelines, Jucikas opened a satchel and pulled out a diagram he’d created. Clickard had told him about the scope of some of the projects, so Jucikas had mapped out a proposed data science pipeline on how to
ingest, cleanse, process, and deploy data from online profiles of people. He had been doing his Ph.D. research on modeling and predicting the behavior of C. elegans roundworms and said that he simply swapped out the worms for people. Jucikas proposed pulling a wide variety of data by building automated data-harvesting utilities, using algorithmic imputations to consolidate different data sources into a single unified identity for each individual, and then using deep-learning neural networks to predict our desired behaviors. We would still need a team of psychologists, he said, to create the narratives needed to change behaviors, but his pipeline served as the first sketch of the targeting system. But what I loved most was that he color-coded it to make the journey look like the London tube map. As he went through his explanation, it was clear that he was perfect for the job.

  So it was that Clickard, Jucikas, and I began working together, and I eventually persuaded Mark Gettleson to join as well. Suddenly I was surrounded by a team of impeccably dressed, blazingly smart, impossibly quirky individuals. And Nix was the ringleader, the grinning, soulless salesman who didn’t understand anything we were doing but wasted no time selling it fast and hard to anyone he thought might pay. He ruled the office with haughty proclamations and crude sexual jokes.

  There was an “anything goes” atmosphere at SCL, perfectly crystallized in a moment that took place a couple of months after we all started. Normally I dress in T-shirts and hoodies, but one afternoon I came into the office after a London Fashion Week event wearing a vibrant burgundy Prada jacket with matching high-waisted trousers and cream-colored Dr. Martens shoes printed with skulls and roses in the style of tattoo flashes. Nix took one look and said, “Chris, what the fuck are you wearing?”

  To which Brent answered, “We fight terror in Prada.”