A Privacy-Preserving Platform for Context-Aware Mobile Social Applications

However, this has privacy implications. Existing applications provide social discovery and social sharing via the cloud. Thus, users have to continuously upload sensitive data collected by their device to the cloud in order to use these services. However, once this data leaves the device, users cannot control anymore how it is used by the cloud. Our aim is to create a platform that can support such applications while giving users control over their privacy.

In this talk, we propose an approach that does social discovery in a device-to-device manner, thereby keeping the sensitive data local to the device. In our approach, we do share content via the cloud, but all content is encrypted (and the cloud does not have the encryption key). Our solution is based on the concept of encounters, which happen every time two devices are within Bluetooth range of each other for at least 2 minutes.

We describe how we used this abstraction to capture meaningful social interactions and how they can be used to enable privacy-preserving, anonymous communication and sharing with individual encounters and groups of people, during of after the physical co-location. Our system also supports selective linkability, which enables users to allow people of their choice to recognize them by name when they encounter, while remaining anonymous to others.

In order to demonstrate the capabilities and usefulness of our system, we built an Android app that visualizes the encounter history on a timeline, correlated with contextual information such as location and calendar entries. It allows users to create groups and share encrypted content with participants.