I know who you are: Deanonymization using Facebook Likes

Abstract

This paper presents a method to deanonymize people using fanpages’ Likes of Facebook users. The strategy shows that information of Likes can be easily crawled from Facebook. Combined with an interactive version of browser-history-stealing it can be used to get identities of users on a website. The attack is possible because of the existence of Facebook’ Likes that can be used as a fingerprint. The claim was tested and discussed with real-world collected data. With the assumption of at least 4 collected Likes per user, 99.91% of them can be deanonymized through the fingerprint of Likes. Apart from that we provide potential solutions for protection of identities in social media.