Capturing Social Networking Privacy Preferences: Can Default Policies Help Alleviate Tradeoffs between Expressiveness and User Burden?

Abstract

Social networking sites such as Facebook and MySpace thrive on the exchange of personal content such as pictures and activities. These sites are discovering that people's privacy preferences are very rich and diverse. In theory, providing users with more expressive settings to specify their privacy policies would not only enable them to better articulate their preferences, but could also lead to greater user burden. In this article, we evaluate to what extent providing users with default policies can help alleviate some of this burden. Our research is conducted in the context of location-sharing applications, where users are expected to specify conditions under which they are willing to let others see their locations. We define canonical policies that attempt to abstract away user-specific elements such as a user's default schedule, or canonical places, such as "work" and "home." We learn a set of default policies from this data using decision-tree and clustering algorithms. We examine tradeoffs between the complexity / understandability of default policies made available to users, and the accuracy with which they capture the ground truth preferences of our user population. Specifically, we present results obtained using data collected from 30 users of location-enabled phones over a period of one week. They suggest that providing users with a small number of canonical default policies to choose from can help reduce user burden when it comes to customizing the rich privacy settings they seem to require.