Cambridge Analytica

There is​ at least one certainty where Cambridge Analytica is concerned. If forty thousand people scattered across Michigan, Wisconsin and Pennsylvania had changed their minds about Donald Trump before 8 November 2016, and cast their votes instead for Hillary Clinton, this small London-based political consultancy would not now be the subject of breathless headlines and Downing Street statements. Cambridge Analytica could have harvested, breached, brain-washed and honey-trapped to their evil hearts’ content, but if Clinton had won, it wouldn’t be a story.

The villains of the piece would no doubt agree with this assessment, but not for very plausible reasons. The exposé conducted by Channel 4 News, with the support of the Observer and the New York Times, captured the now suspended CEO of Cambridge Analytica, Alexander Nix, bragging to someone he believed was a potential client that he’d met Trump ‘many times’ and master-minded the entire Trump campaign strategy. Nix implies that those forty thousand votes were scientifically wrested from Hillary and delivered to Trump thanks to micro-targeted advertising and some especially persuasive messaging. ‘Our data informed all the strategy,’ Nix says, with the swagger of an estate agent reporting that demand for period features is red-hot right now.

It’s true that Cambridge Analytica was recruited to work on the Trump campaign, though not necessarily because of its Machiavellian brilliance. Steve Bannon, Trump’s campaign manager, was on the board of the company at the time, and probably tossed it a contract for some data analysis so as to keep things between friends. When the company first attracted attention in the British media in early 2017, it was reported as also having had close ties to the Leave campaign. In one of several investigative reports on the topic, the Observer’s Carole Cadwalladr wrote in May 2017 that ‘what is happening in America and what is happening in Britain are entwined. Brexit and Trump are entwined. The Trump administration’s links to Russia and Britain are entwined. And Cambridge Analytica is one point of focus through which we can see all these relationships in play.’

Against that heady backdrop, the more recent revelations must register as something of an anticlimax, not least for Cambridge Analytica’s more gullible clients. First, there is no firm evidence that Cambridge Analytica provided consultancy services to any of the major players in the EU referendum of 2016. Nix initially bragged in an article that it had, but confessed to the Digital, Culture, Media and Sport Select Committee in February this year that the article had been drafted by a ‘slightly overzealous PR consultant’. Certainly we should try to understand how ‘Brexit and Trump are entwined,’ but that requires sociological and economic analysis: the process won’t be as simple (or as exciting) as identifying a secret control room.

Second, there is not – and cannot be – any evidence that it swung the election for Trump (by the same token, it isn’t strictly provable that it didn’t), though unsurprisingly the company claims otherwise. This still appears painful for Clinton herself to accept. Interviewed for one of the Channel 4 reports, she speaks of Cambridge Analytica’s ‘massive propaganda effort [which] affected the thought processes of voters’. And yet data analysis is at the heart of modern political campaigning. Clinton, after all, preferred to study data on Michigan from the comfort of her Brooklyn campaign office than actually to visit the state, even as panicking Michigan Democrats pleaded with her to spend time there in the final weeks. If things had turned out differently, there would no doubt have been star-struck puff pieces on the bleeding edge data analytics that were behind the election of America’s first female president.

The scandal is twofold, but neither part really concerns elections. The first is the ‘data breach’, which granted the Trump campaign access to fifty million Facebook profiles without the users’ permission. This occurred by means of the abuse of a simple app called thisisyourdigitallife, the like of which many Facebook users will have come across. Apps such as this appear in a user’s newsfeed as a personality questionnaire, producing relatively banal results which can then be shared with friends. Thisisyourdigitallife was built by Aleksandr Kogan, a psychologist at Cambridge University, to test theories of personality-modelling on the basis of Facebook ‘likes’. Only 270,000 people ever used the app, but it also collected data on their friends – a feature of how Facebook works, rather than of Kogan’s wizardry.

Facebook knew of thisisyourdigitallife, but was under the impression it was geared purely towards academic research. Any user who checked the terms and conditions (and who among us ever does that?) would have thought the same. What nobody knew until earlier this month was that Kogan was passing the data straight to Cambridge Analytica, which was then putting it to work for Bannon. All of this was told to the Observer by a whistle-blower at Cambridge Analytica, Christopher Wylie: the story appeared under the headline, ‘I made Steve Bannon’s psychological warfare tool.’

Various rules were broken. Privacy law assumes that individuals have the right to know what their data is being used for before they agree that someone may collect and keep it. Kogan and Cambridge Analytica were being dishonest and secretive. But while they may not have been true to every letter of their Ts and Cs, and may be in breach of the Data Protection Act, no one, surely, will be surprised to discover that data collected in one arena is put to work in another. Using data in novel (and secretive) ways is virtually the governing principle of the digital economy – what Shoshana Zuboff has termed ‘surveillance capitalism’, and Nick Srnicek calls ‘platform capitalism’.

It’s worth remembering that throughout the 1990s, the internet was viewed as a threat to capitalism as much as an opportunity. Napster was the iconic example. It wasn’t clear where the profits lay, once information was abundant and individual anonymity was the norm. What changed, as Zuboff and Srnicek both explore in different ways, was that the internet began to be treated as a surveillance device of potentially global proportions: cheaper, better or free services were provided on condition that the ‘user’ would be tracked in everything they did and anchored in their offline identity. The fact that most tech giants made – and in Uber’s case still make – vast losses for the first few years of their existence is integral to this strategy. People must be lured into using a service and then kept using it by whatever means necessary; only later is this power converted into revenue.

To suggest that a Facebook user consents to all the ways Facebook uses or might use their data is completely to misrepresent the logic at work here. For that matter, to say that a Guardian reader consents to all the ways the Guardian uses their data (which they deposit every time they visit the website) is to misunderstand the essentially malleable nature of data itself. Its potential value and use emerges after one has collected it, not before.

In the panic surrounding Trump and Cambridge Analytica, this brute capitalist reality has been mysteriously referred to as ‘harvesting’. But if creating an app to collect data without people’s conscious knowledge is ‘harvesting’, then so is a great deal else. By creating free wifi on the London Underground, Transport for London is harvesting data (the wifi network was installed to provide TfL with real-time data on passenger movements). The UK Government Digital Service has harvested data on citizens by manipulating the design of government websites (the common practice of ‘A/B’ testing means different users see different website designs, with data collected on how this affects click-throughs and time spent on each page). Uber harvests data well beyond car journeys (the app continues to collect data on passenger behaviour after a ride has finished, although users can now opt out of this). New digital advertising billboards at Piccadilly Circus are harvesting data (they contain cameras to analyse the facial expressions of people in the crowds passing by).

The second aspect of the recent scandal is grubbier but ultimately less significant. If its own sales pitch is to be believed (an ‘if’ that grows larger by the day), Cambridge Analytica likes to play dirty. Nix and his colleague Mark Turnbull were caught by Channel 4 discussing techniques of honey-trapping, blackmail and counter-intelligence in a manner that owed more to James Bond plots than to psychometrics. Throwaway remarks, that the candidate is just a ‘puppet’ to the campaign team and that ‘facts’ are less important than ‘emotion’, look shady when caught on a hidden camera, but they’re not categorically different from the early ruthlessness of New Labour operators such as Alastair Campbell, Philip Gould and Peter Mandelson. Nor is there any reason to assume that New Labour’s 1990 analogue methods of data analysis – focus group and polling – are less informative or useful than automated psychometrics. As for Nix’s boast that they ‘operate in the shadows’, and his parting shot to the ‘client’ (‘I look forward to building a very long-term and secretive relationship with you’), it’s a wonder the Channel 4 investigator managed to keep a straight face.

So we have a misuse of data, which has rightly attracted the attention of the Information Commissioner’s Office, and some excitable marketing patter, which slips into Mafia fantasy before being swiftly retracted at the first sign of actual danger. The former issue isn’t exactly news: in 2010 the Wall Street Journal found that Facebook apps (such as the one built by Kogan) were routinely collecting information for the benefit of advertisers and internet tracking companies, without users’ consent. Given Facebook’s command of the world’s attention (more than two billion monthly active users, who spend an average of fifty minutes on the site every day), it is inevitable that attention merchants flock to the site in search of the scraps, just as major sporting events attract ticket touts.

Why so much outrage? The Observer should be congratulated for its tenacity on the topic, and this story may, with luck, push us towards a tipping point on the issue of data privacy. But the fascination and shock that Cambridge Analytica is attracting suggests a displacement of horror that really stems from something deeper. Part of that must lie with Trump and Trumpism. A terrible event must surely have been delivered by equally terrible means. Passionate Remainers no doubt feel similarly about Brexit. It is clear that various secretive and underhand forces did intervene in the US election campaign. Thanks to Robert Mueller’s investigation, we know that Facebook sold $100,000-worth of advertising space to Russian ‘troll farms’, and that 126 million Americans may have been exposed to Russian ‘fake news’ over the course of 2015 and 2016. Then there is the FBI’s resurrection of the matter of Clinton’s emails at a critical moment in the election campaign. Whether any of this gets us closer to explaining or understanding Trump’s victory is moot.

Cambridge Analytica looks conveniently like a smoking gun, primarily because it has repeatedly bragged that it is one. Nix and Turnbull do for the events of 2016 what ‘Fabulous’ Fab Tourre, former Goldman Sachs banker, and Fred ‘The Shred’ Goodwin, former boss of RBS, did for the banking crisis of 2008, providing grotesque personalities on which to focus anger and alarm. To hear such men proudly declaring their lack of moral conscience is paradoxically reassuring to the degree that it helps explain the world’s loss of moral direction. But as with the financial crisis, the circus risks distracting from the real institutional and political questions, in this case concerning companies such as Facebook and the model of capitalism that tolerates, facilitates and even celebrates their extensive and sophisticated forms of data harvesting and analysis.

It is telling that two of the greatest ethical scandals to have hit Facebook in recent years both involved academics, the previous one being the ‘emotional contagion’ experiment, in which it transpired that Facebook had altered newsfeeds without consent, as part of a scientific study. Engaging with external researchers means surrendering a tiny modicum of control. Facebook’s willingness to co-operate with academics is already slight, and these scandals will make Mark Zuckerberg wonder if it could ever be worth the bother. Keep all the data in-house and the question of ethics doesn’t arise. The increasing size and scope of these giant platforms gradually eliminates the need ever to share valuable data with anyone else.

It’s sometimes said that data is the ‘oil’ of the digital economy, the resource that fuels everything else. A more helpful analogy is between oil and privacy, a concealed natural resource that is progressively plundered for private profit, with increasingly harmful consequences for society at large. If this analogy is correct, privacy and data protection laws won’t be enough to fight the tech giants with. Destroying privacy in ever more adventurous ways is what Facebook does.

Just as environmentalists demand that the fossil fuel industry ‘leave it in the ground,’ the ultimate demand to be levelled at Silicon Valley should be ‘leave it in our heads.’ The real villain here is an expansionary economic logic that insists on inspecting ever more of our thoughts, feelings and relationships. The best way to thwart this is the one Silicon Valley fears the most: anti-trust laws. Broken into smaller pieces, these companies would still be able to monitor us, but from disparate perspectives that couldn’t easily (or secretly) be joined up. Better a world full of snake-oil merchants like Cambridge Analytica, who eventually get caught out by their own bullshit, than a world of vast corporate monopolies such as Amazon and Facebook, gradually taking on the functions of government, while remaining eerily quiet about what they’re doing.

23 March