Last updated April 1, 2019 at 12:18 pm
Information taken from your friends online can be used to build an accurate profile of you, and predict what you’ll do next.
You may think you’re able to control your privacy on social media, but new research has found that you can be profiled based entirely on the contacts you keep.
Essentially – your friends are ruining your privacy. And you’re ruining theirs.
Published today in the journal Nature Human Behaviour, researchers from the University of Adelaide and the University of Vermont have found that even where people have deleted their accounts they can be profiled from the information that can be drawn from their friends’ posts.
Amalgamating data from over 30 million Twitter posts by nearly 14,000 users, they discovered that a person’s future tweets, activities and interests could be predicted from just eight to nine of their contacts with 95% accuracy.
That accuracy was comparable to using the target’s own profile data.
Your information is embedded with friends
“It shows that information about you is embedded in your interactions with friends,” Lewis Mitchell from the University of Adelaide told Cosmos Magazine.
“Just like if I overhear one side of a phone conversation it tells me something about the person on the other end of the line, so do your friends’ social media posts tell me something about you.
“And if you delete your account, it doesn’t necessarily help, because you can still in principle be profiled from the digital traces left by your interactions with friends.”
This could include private information such as friendships, religion, sexual preferences and even a person’s location, says David Garcia from the Medical University of Vienna, in a related editorial.
Indeed, people were able to be profiled even if they had never signed up to a social media platform, using information exclusively from friends’ accounts.
Facebook CEO Mark Zuckerberg has admitted the platform does exactly this, collecting information on people who are not users.
There are, however, limitations about how well this research translates in the real world. In particular, the predicted information can quickly become outdated as the social media platform and people themselves change. If someone were to close their social media account it could even affect their contact’s online behaviour, if for no other reason that the contact is no longer interacting with the target person. That change in interaction itself could affect the model’s accuracy.
Concerns over privacy in an online society
Still, the research raises serious questions about privacy and how people’s choices are tracked and embedded in social networks.
There is the possibility that organisations, governments, or private companies could build an accurate profile of someone, and predict their decision making, entirely from their friends.
And in a world where online media is increasingly being used to coordinate political and social demonstrations and movements, it raises concerns about misuse of that information.
“This all means that there is no place to hide in online social networks like Facebook and Twitter,” Mitchell says.
“There are benefits from being able to predict behaviour. Social media platforms use this principle to target information so that you receive posts that you are interested in.
“But of course, there is a dark side as well, such as the potential for the creation of ‘filter bubbles’. For instance, in a political debate, people may be only exposed to one type of information and may not receive any opposing views,” warns Mitchell.
“You alone don’t control your privacy on social media platforms,” conclude the authors.
“Your friends have a say too.”