How can algorithms know us, when we hardly know ourselves?
Social media networks like Facebook seem to know so much about us in unnerving detail, but understanding why we press 'Like' is still beyond even the cleverest code.
The oft quoted expression, “If something is free, you are the product” is no more true than when it comes to social media and Internet searching. Today we see more than ever how much can be known about us through our online activities. When it comes to algorithms there are no accidents. That lampshade you googled yesterday appears in your Facebook feed today along with other soft furnishings that fit your taste. Your newsfeed is curated for your personal style, and the profile based on your online activities know you better than you know yourself. Or do they?
Human beings are complex, and social media only appeals to parts of who we are. In the past, I have referred to our social media identities as a digital extension of the self - but it would be more accurate to call it a digital extension of the partial self.
Social media is like a virtual public environment, one that accesses parts of the self that are “outward leaning” - often at the expense of the parts that either lean inward (our private selves) or that we feel reluctant to share. It’s not so much that what is known about us via our online personas isn’t true - it’s that it’s partial.
A 2017 article in Motherboard noted just how much can be known about us simply from what we like on Facebook – and it’s a lot. Researchers were able find out by looking at what people liked on Facebook and then comparing that to psychometric tests they’d taken earlier on that very same platform. This test reveals what psychologists call “the big five personality traits” which includes characteristics like how open, conscientious, agreeable, extroverted or neurotic you are. These traits are considered to be the most-highly validated personality measures in psychology.
It is reported that millions of people took these tests for a bit of fun on Facebook. With such massive amounts of data to work with and refine, their results were frighteningly predictive of personal choice and behaviour. According to Motherboard, the researchers found that a mere 70 Facebook likes provided enough information to know as much about you as your close friends: with three hundred, they could know you as well as your partner. Pretty freaky stuff – and it seems there’s little doubt that they could indeed discern your tastes, your sexuality, and your political persuasion and a great deal more based on a series of thumbs-ups posted over time.
But let’s think about this a little more deeply. Of course these “big five” have a lot to do with how we think about of ourselves and others. After all, we can tell if we’re conscientious or not if we’re generally able to meet deadlines, or by the state in which we keep our desks. We identify others by whether they’re agreeable (nice) or a bit spikey and some of us deal with challenges in a calm and collected way while others are more likely to worry – an indicator of how neuroticism. But how well do these measures tell us about how it is to be who we are?
The Myers-Briggs Type Indicator (MBTI) is less validated by psychologists, but often more appreciated by those who actually take it. That’s because it’s based on the kind of person you are within different contexts. For example, do you make decisions based on intuition or by rationally analysing the facts? Do you get energised by large groups of people, or are you more into an intense tête-a-tête? Chances are you might be both or either, depending on the situation.
More like this
The way you “attach” to others in relationships is another kind of personality type. When it comes to relationships are you slow to say, “I love you”, do you dive straight in, passionately and intensely, do you get insecure, or you swing wildly between?
The MBTI is based on the Jungian ideas whereas Attachment Theory comes by way of the British psychoanalyst John Bowlby. These are both “psychodynamic” theories, which means they appreciate the power of the unconscious – a powerful force, but difficult to measure.
There is little doubt that how we present ourselves via social media gives away a huge amount of information about us – but in a narrow frame. We share only what we are prepared to, and we do so knowing that we’re being watched. What’s going on below the level of the “share” is much more complex. After all, how well do we even know ourselves? How do we make the choices we do, and why do we find ourselves behaving in ways that perplex us?
Many people spend years in therapy trying to work out these details – to unpeel layers of history and experience to understand how and why we are who we are. Sharing a humorous gif, liking a certain brand, or sharing your outrage about a current event does indeed tell people something about you – but only part of the story. At Stillpoint Spaces London, we have created a psychological “Lab” where we try to better understand more of the story by looking at contemporary life through a psychological eye.
Today, it is even more imperative to examine these issues closely for the very preservation of our liberal democracy. A recent investigative report by Carol Cadwalladr at The Observer reveals how millions of dollars of investment are going into Internet architecture and being used to manipulate the way in which we access news with the aim of changing our opinions and influencing our choices.
Approaching these issues from within a narrow “consciousness bias” won’t be enough to truly understand the important issues here. At the same time, the idea that we can be “known” by what we like on Facebook also needs to be critically reappraised. Sure, you may easily work out how someone is likely to vote, what brand they are likely to buy, and even if one will soon be a new mother. But algorithms and psychometrics won’t quite tell you the nature of someone’s values, why they’re attracted to one brand and not another, and how it feels to be expecting. One of life’s great adventures is discovering who we are and getting to really know others: even the best algorithms won’t do that.