Abstract

The widespread use of smart mobile devices has fostered the development of a variety of successful data-sharing mobile applications. The data that users disclose to untrusted entities (for example, through location-based services) exposes aspects of their private life, which is not apparent at first but can be inferred from the revealed data. People are notoriously bad estimators of risks, including privacy risks. Moreover, due to various cognitive biases, lack of enough information, and the complexity of the decision problem, it is difficult for users to make the optimal decision about whether to reveal or obfuscate their information and, if necessary, how to obfuscate it. In this talk, I address the problem of protecting users’ privacy in data-sharing mobile applications, with the focus on location-based services. I propose strategic algorithms and intelligent tools to automatically assess the users’ location-privacy level, and to find the right balance between revealing and hiding their data. I will present the Location-Privacy Meter (LPM) tool, that we developed to quantify location privacy, as well as our work on optimal defense mechanisms against location-inference attacks.