Faking contextual data for fun, profit, and privacy

Abstract

The amount of contextual data collected, stored, mined, and shared is increasing exponentially. Street cameras, credit card transactions, chat and Twitter logs, e-mail, web site visits, phone logs and recordings, social networking sites, are all examples of data that persists in a manner not under individual control, leading some to declare the death of privacy. We argue here that the ability to generate convincing fake contextual data can be a basic tool in the fight to preserve privacy. One use for the technology is for an individual to make his actual data indistinguishable amongst a pile of false data. In this paper we consider two examples of contextual data, search engine query data and location data. We describe the current state of faking these types of data and our own efforts in this direction.