The most interesting question about 2016 is not why the Brexit result and Trump happened, but whether historians will regard both as incidental; whether this will go down as the year democracy revealed itself unworkable in the age of the internet – in which reality, already engaged in a life-or-death struggle with inverted commas, finally gave way to “alt-reality”.
The results of these votes were shocking, but not surprising. The rules of capitalism have been gamed by the ruling kleptocracy and a lot of working people are angry. No mystery there. In the past week, however, the collective postmortem – on the left and right of politics – has focused on a concern with far greater long-term impact: the accidental or deliberate propagation of misinformation via social media. Many millions of people saw and believed fake reports that the pope had endorsed Trump; Democrats had paid and bussed anti-Trump protesters; Hillary Clinton was under criminal investigation for sexually assaulting a minor. About the only accusation not levelled at Clinton was implication in the murder of JFK, and that was because Trump had already used it against his Republican primary rival Ted Cruz. If democracy is predicated on reliable information, it’s in serious trouble right now.
Very few people saw this coming. Back in the 1990s, at the height of the dotcom boom, the internet pioneer Josh Harris tried to sound a warning – but at that early utopian stage, when the web was assumed to be decentralising, democratising, enlightening, almost no one understood what he was saying. Later, in 2002, George W Bush’s own Voldemort, his deputy chief of staff Karl Rove, chided a reporter by saying: “People like you are in what we call the reality-based community. You believe that solutions emerge from judicious study of the discernible reality. That’s not the way the world really works any more.” The gnomic taunt caused more bemusement than consternation at the time, but Rove was ahead of the game.
How did we get here – and was/is this situation inevitable? Much has been written about the arrival of a “post-truth” era, in which facts become secondary to feeling; expertise and vision to ersatz emotional connection. Nazi Germany shows that this is not new, but the internet-driven efficiency with which it can be manipulated is.
Also new and specifically net-enabled is something I call “retro-truth”, which marks a deep existential and ethical shift for humanity, in which a proposition is judged not by whether it is true or false when stated, but whether it has the potential to become true – whether it contains the possibility of truth, like energy waiting to be released from an atom. Both developments stem from the speed and stealth with which information now travels and they challenge society’s most fundamental ethical underpinnings. In their shadow, traditional notions of truth and falsity merge into each other, allowing anything to be said with impunity.
There’s a technocratic side to the information problem as well. At the Miami book fair last week I met Cathy O’Neil, author of Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy. O’Neil was a teen maths prodigy-turned-professor who left academia for a hedge fund before realising post-2008 that “mathematics, once my refuge, was not only deeply entangled in the world’s problems, but also fuelling them”.
O’Neil’s book describes the malign effects of the “opaque, unquestioned, unaccountable” algorithms – known as algos – that began in finance but now mediate every aspect of our lives, including what we see in our Facebook news feeds. Her overarching message is that nothing about these algorithms is neutral, with editorial decisions unavoidably written into the software despite being presented as automated and therefore objective. “There’s always a measure of success for any algorithm,” she told me, “and the general rule is that if you don’t know what that is, it’s probably profit.” Algos also punish the poor, she argues, because the rich tend to have their affairs managed by people rather than cheap monolithic software.
Back in July, post-Brexit vote, this paper’s editor, Katharine Viner, detailed the ways in which net organisations such as Facebook and Google have undermined traditional journalism. One of the main drivers of this process is a click-based revenue model, in which algorithms prioritise items in news feeds based on how likely individual users are to “engage with” (ie click on) them – and thus be exposed to more ads. Whether these items contain carefully researched or fabricated material is of no concern to the algos: in fact, false, sensationalist stories that bolster existing prejudices are more likely to draw clicks than sober analyses that challenge assumption. With misinformation being incentivised in this way, who could be surprised when Buzzfeed found a group of young Macedonians copying the most outlandish fabrications to more than 140 specially created pro-Trump websites and sexing up the headlines to gain clicks and go viral on Facebook?
This is important, because a recent study by the Pew Research Centre found a majority of American adults using Facebook as a source of news (which means Britain is sure to follow). Facebook CEO Mark Zuckerberg has been resistant to the notion that his company, social media, or the web in general are undermining democracy (“a pretty crazy idea”), even after dozens of his own staff formed a covert taskforce to address the problem post-election. It’s easy to see why he bridles too, because if he accepts the truth that his algorithms function no more objectively than a human editor, then he bears responsibility for their choices. And once he does that, he allows the equally obvious truth that Facebook, whether it wants to be or not, is now a media organisation and must vouch for the information it disseminates.
Among the most pernicious myths of our time is that the functioning of the web is neutral and immutable; that it has evolved of its own ethereal logic, like a galaxy, and can’t be changed or stopped. The web we have is one of many we might have had – indeed, when Tim Berners-Lee released its protocols in the early 1990s, some seasoned net-heads were appalled by its primitive mimicry of paper, lack of deference to copyright and so on.
What can be done? O’Neil champions the idea of a Hippocratic oath for the “quants”, people who create and nurture algorithms, and of regular algorithmic audits to assess their effects. More fundamentally, it would be easy for us to decide as a society and legislate accordingly that Facebook et al should be treated as media organisations, and held accountable for the information from which they profit. And it’s important that we act now, because the same truth trolls who disgraced the US election are turning their fire on the socially minded entrepreneur Elon Musk, perhaps fearful that he might go into politics – and rest assured that it won’t end there.
One day, I suspect, we will look back in disbelief that we let the net-induced friction on civil society reach this pitch, because if we didn’t know before, we know now that our stark choice is between social networks’ bottom line and democracy. I know which I prefer. Or will we let the algos decide?
