Is Facebook/Instagram spying on you?
Spoiler alert: no. They don’t need to and really can’t be bothered. They already know everything about you. Everything they need to. And everything else.
Are they listening to my conversations?
How many times have you had a conversation with someone about something only to have it served up in an advert on Facebook or Instagram the next day? It’s creepy right? My brother had a few friends over and at some point someone mentioned a rare Japanese whisky they had heard of. The next day my brother received adverts about that exact whisky. How do they know? It feels like they must be listening in to our private discussions, teasing out product mentions and selling adverts to urge us to buy. It has been noticed by many people so many times. e.g. Facebook ‘listening’ claim denied by professor.
A mobile security company, Wandera, set up some tests with mobile phones in a controlled environment and proved that there was no recording and transmission back. It also wouldn’t make sense for a company to listen in, not just because of the privacy violations, but because the resource requirements would be massive. It would cost an enormous amount, it’s not massively scalable and the cost wouldn’t be recouped with a few adverts. In short, it’s unlikely.
What is happening is data interpolation of who you’re with and what your interests are. Wherever you take your mobile phone, it is pinging off local Wi-Fi spots, even if you’re not logged in to them, to improve the accuracy of your location. This is how “find my iPhone” and other location based services work. And yes, you have agreed to this tracking in the terms and conditions you scrolled past and accepted. With this data (and yes, you allowed them access to it too), Facebook can determine where you and more crucially, who you are with.
The other data element that Facebook uses here is web caches or cookies, i.e. your browser and search history. Everything you look at on the internet, most of the apps you use will store data on your usage and your interests. It is then not a huge leap to say that if Anne and Jude spent time together in the same place last night, like Anne’s house party, then they will have similar interests. And based on Jude’s personality profile (more on this later) and Anne’s recent browser history (i.e. things she has been interested in recently), Facebook can cross the two and serve up ‘relevant’ adverts to Jude.
Where it starts to feel even more freaky is the post-meet data cross. At the party Anne and Jude discuss a remote and relatively obscure surfing location in Morocco that one heard about earlier from a friend but neither has looked into yet. On her taxi ride home, Jude searches for it and spends some time looking at travel options and beginner courses. The next morning, Anne has it in her mind but in checking her Instagram, receives adverts about the exact location they were talking about, the Panoramas in Taghazout, with some offers for holidays and surfing courses! Whilst she feels like they must have been listening in, it is an algorithmic sleight of hand that leverages location data and search history information, which can be scaled across the billions of Facebook users.
But are they listening to my THOUGHTS??
How many times have you just thought about something and had it served up in an advert a few hours later? This feels super creepy. How do they know what I’m thinking? Are they in my head?
There are two drivers behind this experience: (a) Facebook know everything about who you are and what you like doing and (b) they have powerful predictive algorithms that can guess to a very high degree of accuracy what and how you will think and behave. In many respects, yes, it is super creepy.
Unpacking these elements: (a) they know everything about who you are. This comes down to people unknowingly giving their data away.
Did you ever have the craze of “what’s your porn-star name?” It’s your first pet’s name with your mother’s maiden name. And people put their answers online and said, “Oh it’s so funny! Spinksy Clifford! Ha ha!”. But this was a classic phishing exercise as a lot of passwords contain animal names and lots of security questions ask for your mother’s maiden name. Without thinking, these details were proffered up on the internet in the name of fun.
You may remember around ten years ago there were a lot of quizzes on Facebook: what kind of personality are you? The amount of personal data freely given away in these exercises was immense and provided a huge database of millions of peoples’ personality types which could then be correlated with their internet behaviour. Using this data, the Cambridge psychometrics team developed models that allowed them to determine to a 90% accuracy your personality type with just 10 Facebook likes — not comments, just likes. With 100 likes they could get to 99% accuracy on your personality type and behavioural tendencies and with 300 likes, it is said they would know you better than your closest friend, your spouse or anyone.
In the excellent 1997 film, ‘The Game’, with Michael Douglas, the main character is psychologically profiled then driven through a series of events that provoke him to react in a number of ways, culminating in him shooting his brother and then throwing himself off a roof. The whole thing is a set-up and everyone comes out unscathed, but it begs the question how far our actions can be manipulated by someone with deep knowledge of our inner thoughts and feelings.
How does the predictive piece work? What was known as “big data” and is now called “artificial intelligence” is for the most part highly powered algorithms that can take huge streams of disparate data and find unique correlations between them. These algorithms can then make highly accurate predictions of behaviour given certain inputs. So if you know Anne (age X, socio-economic status Y, income Z, interests A, B, C) reacted in a positive way to an advert then you can deduce a high probability of how Jude, of similar profile, may respond to that advert.
Should I be worried by what they know about me?
Ultimately Facebook wants to make money by selling advertising space to marketeers. The better they can do that, the more money they make. The knowledge they have of their audience through their vast data sets is why they are valued at over half a trillion dollars. All this data they are using to better target adverts with greater accuracy with better outcomes. But who has access to this data and who can leverage it and to what ends is a cause for concern.
The EU has started to worry (perhaps too late) about the extent of this data collection and the ways in which it may be used. The compelling TV movie ‘Brexit: The Uncivil War’ highlighted how manipulation of this data and its insights had been used by Vote Leave with AggregateIQ and Cambridge Analytica to influence the outcome of the UK’s referendum vote on EU membership. The Netflix documentary ‘The Great Hack’ points out the same approach also used in US elections to drive voters’ behaviours to elect President Trump.
Ultimately what is at stake here is not the ‘creepy’ individual adverts you receive based on what you’ve discussed with a friend, but the erosion of democracy through these highly targeted influencing activities appropriated by self-interested parties and thence a corrosive degradation of our liberty. Ironically, remaining in the EU may have helped the UK in overcoming or moderating some of these nefarious digital influences. What the future holds, only the algorithms know, or more pertinently, those paying for them.
