Forget about your privacy – you’ve got friends!
Ever heard about the shadow profile hypothesis? The idea is around since we know that companies like Facebook and Google (and so many more) collect private data of their users on a large scale.
What if they went beyond that? If they would not stick to their users but collect data about the “friends” of their users as well, regardless if they have got their own Facebook / Google / whatever else / account or not?
People love gossip.
Companies love the information in gossip.
This form of information gathering uses the gossip principle.
Take a typical conversation as it could take place every day everywhere in the world. “Have you heard? Susie is going out with Pete now! … Of course you know Susie: the blonde girl we always meet at the gym on Friday. Pete? He is her trainer!” …
With such bits of second hand information you would make a pretty good first guess about some aspects of Susie: For sure you’d know that she is young, blonde and a woman, most likely heterosexual and sportive. Her new trainer-friend Pete is sportive for sure.
In normal life you probably won’t pay much attention to the chitchat of strangers, although you could easily learn a lot about the gossipers – their interests, friends, their living circumstances – and their unseen friends. But a similar kind of information exchange on the level of a network with millions and millions of participants would allow not only to paint a very accurate picture of the people within the network – and thus to place highly personalized advertising.
It would allow to create a profile of people outside the network as well: a so called shadow profile.
Shadow profiling? Proof of concept
CSH’s complexity scientist David Garcia shows how easy it is to create such shadow profiles with the secondhand information given by friends.
Recently he used recovered data from Friendster, a social network that got wiped out by the appearance of Facebook, to show that the communication of friends can improve the estimation of the sexual preference of people who were not on Friendster themselves.
Now, David and colleagues tested the shadow profile hypothesis for the first time with data from Twitter.
The researchers chose more than 1,000 existing twitter users and their friends, who generated more than 150 Million tweets, to successfully track down the location and to gain biographical insights from Non-Twitter users.
“We don’t know for sure that shadow profiles exist”, says David, “but there is proof enough how easy it is to create them. This mere possibility should be a huge privacy concern for politics – and for each of us.”
Privacy is not an individual’s decision anymore
David’s and similar work of others show that online privacy is not an individual choice anymore. “Without oversight or collective control mechanisms”, his paper published in the journal EPJ Data Science states, “individuals have little power to ensure that they are not being profiled without their knowledge or consent.”
Meaning? Being careful or careless with personal information – we are no longer the masters of our data. The very moment we’ve got friends communicating about us on social media we are an open book to data leeches.
Human and Machine Interaction in Complex Production Systems [German & English]
TU Graz_research, Jun 15, 2018