RT @IWantPrivacy: Widespread Violation of Privacy Settings in the Twitter Social Network

Abstract

Twitter is a social network that focuses on creating and sharing short 140 character messages know as tweets. Twitter's sole privacy policy is a binary option that either allows every message a user creates to be publicly available, or allows only a user's followers to see posted messages. As the Twitter community grew, conventions organically formed that allow for rich expressiveness with only 140 characters. Repeating what someone else says is called retweeting; this behavior facilitates the spread of information and commentary in real-time. We have performed a large-scale collection of data from the Twitter social network by means of the publicly available application programming interface (API) they provide. Our data set contains over 2.7 billion messages, 80 million user profiles, and a 2.6 billion edge social network. We analyze these data and uncover the growing trend where users defeat Twitter's simple privacy mechanism of "protecting one's tweets" by simply retweeting a protected tweet. We have shown through quantitative and qualitative analysis that these privacy-violating retweets are a growing problem. More than 4.42 million tweets exist in our corpus that expose protected information. As Twitter gains popularity over time, we see an increasing trend in the number of privacy-violating retweets. Although there are many users who are unaware that their private tweets are being broadcast to the public, there are some who are aware of this problem.