Facebook’s collection of data makes it one of the most influential organisations in the world. Share Lab wanted to look “under the bonnet” at the tech giant’s algorithms and connections to better understand the social structure and power relations within the company.
A couple of years ago, Vladan Joler and his brainy friends in Belgrade began investigating the inner workings of one of the world’s most powerful corporations.
The team, which includes experts in cyber-forensic analysis and data visualisation, had already looked into what he calls “different forms of invisible infrastructures” behind Serbia’s internet service providers.
But Mr Joler and his friends, now working under a project called Share Lab, had their sights set on a bigger target.
“If Facebook were a country, it would be bigger than China,” says Mr Joler, whose day job is as a professor at Serbia’s Novi Sad University.
He reels off the familiar, but still staggering, numbers: the barely teenage Silicon Valley firm stores some 300 petabytes of data, boasts almost two billion users, and raked in almost $28bn (£22bn) in revenues in 2016 alone.
And yet, Mr Joler argues, we know next to nothing about what goes on under the bonnet – despite the fact that we, as users, are providing most of the fuel – for free.
“All of us, when we are uploading something, when we are tagging people, when we are commenting, we are basicallyThe data our interactions provide feeds the complex algorithms that power the social media site, where, as Mr Joler puts it, our behaviour is transformed into a product.
Trying to untangle that largely hidden process proved to be a mammoth task.
“We tried to map all the inputs, the fields in which we interact with Facebook, and the outcome,” he says.
“We mapped likes, shares, search, update status, adding photos, friends, names, everything our devices are saying about us, all the permissions we are giving to Facebook via apps, such as phone status, wifi connection and the ability to record audio.”
All of this research provided only a fraction of the full picture. So the team looked into Facebook’s acquisitions, and scoured its myriad patent filings.
The results were astonishing.
Visually arresting flow charts that take hours to absorb fully, but which show how the data we give Facebook is used to calculate our ethnic affinity (Facebook’s term), sexual orientation, political affiliation, social class, travel schedule and much more.working for Facebook,” he says.
One map shows how everything – from the links we post on Facebook, to the pages we like, to our online behaviour in many other corners of cyber-space that are owned or interact with the company (Instagram, WhatsApp or sites that merely use your Facebook log-in) – could all be entering a giant algorithmic process.
And that process allows Facebook to target users with terrifying accuracy, with the ability to determine whether they like Korean food, the length of their commute to work, or their baby’s age.
Another map details the permissions many of us willingly give Facebook via its many smartphone apps, including the ability to read all text messages, download files without permission, and access our precise location.
Individually, these are powerful tools; combined they amount to a data collection engine that, Mr Joler argues, is ripe for exploitation.
“If you think just about cookies, just about mobile phone permissions, or just about the retention of metadata – each of those things, from the perspective of data analysis, are really intrusive.”