Fitness trackers and wearable devices: How to prevent inference risks?

Wearable and personal devices are becoming more and more part of people’s everyday lives. These devices produce enormous amount of personal data which are handled by third parties as authorized by the user. However, such third parties may be able to infer sensitive information using the collected personal information. In this paper we present a case study based on fitness trackers and we sketch our model for privacy management and inference prevention. For this study, we built a Bayesian Network and used it to compute the risk of inferring unknown data. Using the simulated case we show the feasibility of inferring some private data from a set of personal data available to a third party as authorized by the user (i.e., sensor data and profiling data provided by the user while registering for the service). This paper provides a step towards the open issues of privacy and security management in the field of ubiquitous devices.