Exploiting social networks to provide privacy in personalized web search

Web search engines (WSE) have become an essential tool for searching information on the Internet. In order to provide personalized search results for the users, WSEs store all the queries which have been submitted by the users and the search results which they have selected. The AOL scandal in 2006 proved that this information contains personally identifiable information which represents a privacy threat for the users who have generated it. In this way, AOL released a file containing twenty million queries made by 658,000 persons and several of those users were successfully tracked. In this paper, we propose a P2P protocol that exploits social networks in order to protect the privacy of the users from the profiling mechanisms of the WSEs. The proposed scheme has been designed considering the presence of users who do not follow the protocol (i.e., adversaries). In order to evaluate the privacy of the users, we have designed a new measure (the profile exposure level (PEL)). Finally, we have used the AOL’s file in order to simulate the behavior of our scheme with real queries which have been generated by real users. Our tests show that our scheme is usable in practice and that it preserves the privacy of the users even in the presence of adversaries.

► We propose a protocol that protects the privacy of the users of web search engines.
► The proposed scheme addresses the presence of users who do not follow the protocol.
► We have designed a new measure to evaluate the privacy of the users.
► We have used real queries from real users to simulate the behavior of our scheme.