The National Science Foundation (NSF) awarded a three-year, $474,000 grant this month to Evimaria Terzi, professor of computer science at Boston University’s College of Arts & Sciences. The grant will fund research to develop tools to protect the privacy of social media users. Terzi is the principal investigator for the project.
The grant addresses a growing need of people who use social-networking sites (such as Facebook, MySpace, and LinkedIn) and other online collaborative tools where sensitive or personal information is posted and shared. Although many sites already allow users to control access to their data, studies have shown that most users find existing access-controls too complicated and therefore don’t bother to use them. The goal of this project is to help collaborative and social-media users gain control of their data by providing intuitive, easy-to-use security features.
Terzi’s approach is based on three key components: assisted specification, feedback, and refinement recommendations. To assist users in initially specifying access-control policies for their data, the project will develop a “privacy wizard,” which employs data mining and machine learning methods, including active learning, to construct an accurate policy with minimal input from the user. Feedback regarding existing privacy settings will come from two additional approaches: aggregate scores and visualizations. Aggregate scores explain to the user how her settings differ from those of other users. Visual feedback will consist of a visualization tool that will allow the user to visually infer the differences between her settings and the settings of their friends and other users. Users’ settings are then refined via a recommendation system that uses users’ reaction to feedback in order to recommend settings that allow the user to have the appropriate level of social exposure.
Online collaborative tools and social media offer great promise in a number of arenas, including business, medicine and education. However, the absence of usable privacy and access controls has slowed the adoption of these tools for serious applications. The results of this project, which will be disseminated via prototype implementations and research publications, will help social media technology achieve its full potential.