Incognitus: Privacy-Preserving User Interests in Online Social Networks

Abstract

Online Social Networks have changed the way we reach news and information. An increasing number of people use social networks not only for communicating with friends and colleagues but also for their daily information needs. Apart from providing the users with personalized information in a timely manner, this functionality may also raise significant privacy concerns. The service provider is able to observe both the Pages a user is subscribed to and her inter- actions with them. The collected data can form a detailed user profile, which can later be used for several purposes; usually beyond the control of the user. To ad- dress these privacy concerns, we propose Incognitus: an approach to allow users browse Pages of OSNs without disclosing their interests or activity to the service provider. Our approach provides (i) a incognito mode of operation when browsing privacy-sensitive content. In this isolated, offline mode no tracking mechanisms can monitor the users behavior and no information can be leaked to the provider. At the same time, (ii) by using an obfuscation-based mechanism, Incognitus reduces the accuracy of the service provider when monitoring the interests of a user. Early results show that Incognitus has minimal bandwidth requirements and imposes reasonable latency to the users browsing experience.