This project will develop Content-Based Access Control (CBAC), a system for access control on content sharing sites. The development of this system involves three research efforts. First, the research will develop and extend techniques for the automatic content analysis, to categorize entries on content-sharing sites and identify potential readers. Second, it will implement a CBAC system for creating and enforcing content-based access control policies. Third, it will develop techniques for adapting these policies automatically based on observation of users' content, behavior and context. This research will be evaluated in the context of PLoG, a Privacy/Policy aware blogging system.
The popularity of social networking sites, blogging, and other content-sharing sites has exploded, resulting in more personal information and opinions being available with less access control than ever before. However, many sites provide only the most rudimentary of access control mechanisms: a document can be either completely private or completely public. Even the best provide only general role classifications such as private, friends-only, and world-readable, making it impossible for a user to place multiple facets of his or her life into a single, controlled repository. The consequences of poor access control mechanisms have been well-documented. For example, some bloggers have lost their jobs or become victims of stalking.
This interdisciplinary project will develop an access control system that is easy for users to understand and administer. To support the system, it will develop new and extended techniques for automatic categorization of text and multimedia objects, access control policy authoring and maintenance, and automatic adaptation of access control policies. These techniques will enable the millions of people who contribute to content sharing sites to better control the information they share with others.
