Going once, going twice… shared!

Sharing content on social networks can create privacy concerns: others could easily share your picture or other personal information without your permission. PhD candidate Onuralp Ulusoy designed a mechanism to tackle these issues. He used an auction-based approach, where users place bids to reach a collaborative outcome. He also designed a digital assistant to do the bidding for you. Ulusoy defended his thesis on 7 December.

Imagine two similar pictures featuring you and three colleagues, taken during a business outing. As a group, you want to post one of the pictures online, but you disagree on which one it should be. While you prefer to share the picture in which you don’t have your eyes closed, the others prefer the second picture. What picture should be shared?

Placing a bid

Ulusoy came up with a mechanism that could resolve this predicament. He designed a collaborative approach that preserves privacy of social network users, while at the same time still enabling users to share content. The approach is auction-based: all users who might have their privacy affected, i.e. you and your colleagues in the pictures, have a say in the final decision. They do so by placing bids with a digital currency. The more they bid, the heavier it weighs in the final outcome. So, if you have a mild preference for one picture over another, you can place a low bid, but if you are a hundred percent sure you do not want a picture of yours being shared, you place a high bid.

This system forces users to really weigh their options and think consciously about what they do and do not want to share

Under-sharing

But what if you never want any of your pictures to be shared? Placing a high bid reduces the amount of currency you have. So, when another not so nice picture of you shows up, placing a high bid is not an option anymore, and the picture could be shared against your will. Ulusoy emphasizes that the system is designed to seek a middle ground between under-sharing and over-sharing. “Social media is a content-driven environment. It exists by virtue of users sharing content – as long as it is not really private, of course”, he says. “This system forces users to really weigh their options and think consciously about what they do and do not want to share.” He adds: “Currently, any of your pictures can be shared anywhere without you knowing it. With this system, at least you get notified and have a say in the final outcome.”

Privacy assistant

Ulusoy also created personal privacy assistants to place bids on a user’s behalf. These virtual assistants learn the privacy requirements of the users they represent, and identify emerging privacy norms on social media. These are norms that change depending on how people in general behave on social media. For instance, it is not prohibited to share pictures of people being drunk in a business outing, but it could be an emerging pattern that sharing these kinds of pictures is getting less and less desirable. The privacy assistant will incorporate this and existing norms in its bidding procedures.