The highly interactive nature of interpersonal communication on online social networks (OSNs) impels us to think about privacy as a communal matter, with users' private information being revealed by not only their own voluntary disclosures, but also the activities of their social ties. The current privacy literature has identified two types of information disclosures in OSNs: self-disclosure, i.e., the disclosure of an OSN user's private information by him/herself; and co-disclosure, i.e., the disclosure of the user's private information by other users. Although co-disclosure has been increasingly identified as a new source of privacy threat inherent to the OSN context, few systematic attempts have been made to provide a framework for understanding the commonalities and distinctions between self-vs. co-disclosure, especially pertaining to different types of private information. To address this gap, this paper presents a data-driven study that builds upon an innovative measurement for quantifying the extent to which others' co-disclosure could lead to actual privacy harm. The results demonstrate the significant harm caused by co-disclosure and illustrate the differences between the identity elements revealed through self-and co-disclosure.

Today, one can find from the web vast amount of information about an individual. Specifically, such information can be classified into two categories, virtual and real-world identities. This paper addresses a novel problem of linking these two types of identities based on information publicly available on the web. We start by studying how one can link virtual identities (i.e., user profiles) at Twitter with real-world identities at Whitepages.com (containing personal information such as name, age, relatives, etc.). We demonstrate that a substantial portion (at least 0.17%) of Twitter users in the U.S. can indeed be potentially linked to their real-world identities through information available at Whitepages.com, revealing sensitive personal data. We discuss the implications of such identity linkages on both individual privacy and law enforcement, and also point out the future studies required in this topic. Keywords-Twitter; virtual identity; Whitepages; real world identity; identity linkage; identity elements978-1-4799-9889-0/15/$31.00 ©2015 IEEE

No abstract

Among the existing solutions for protecting privacy on social media, a popular doctrine is privacy selfmanagement, which asks users to directly control the sharing of their information through privacy settings. While most existing research focuses on whether a user makes informed and rational decisions on privacy settings, we address a novel yet important question of whether these settings are indeed effective in practice. Specifically, we conduct an observational study on the effect of the most prominent privacy setting on Twitter, the protected mode. Our results show that, even after setting an account to protected, real-world account owners still have private information continuously disclosed, mostly through tweets posted by the owner's connections. This illustrates a fundamental limit of privacy self-management: its inability to control the peer-disclosure of privacy by an individual's friends. Our results also point to a potential remedy: A comparative study before vs after an account became protected shows a substantial decrease of peerdisclosure in posts where the other users proactively mention the protected user, but no significant change when the other users are reacting to the protected user's posts. In addition, peer-disclosure through explicit specification, such as the direct mentioning of a user's location, decreases sharply, but no significant change occurs for implicit inference, such as the disclosure of birthday through the date of a "happy birthday" message. The design implication here is that online social networks should provide support alerting users of potential peer-disclosure through implicit inference, especially when a user is reacting to the activities of a user in the protected mode.

Among the existing solutions for protecting privacy on social media, a popular doctrine is privacy selfmanagement, which asks users to directly control the sharing of their information through privacy settings. While most existing research focuses on whether a user makes informed and rational decisions on privacy settings, we address a novel yet important question of whether these settings are indeed effective in practice. Specifically, we conduct an observational study on the effect of the most prominent privacy setting on Twitter, the protected mode. Our results show that, even after setting an account to protected, real-world account owners still have private information continuously disclosed, mostly through tweets posted by the owner's connections. This illustrates a fundamental limit of privacy self-management: its inability to control the peer-disclosure of privacy by an individual's friends. Our results also point to a potential remedy: A comparative study before vs after an account became protected shows a substantial decrease of peerdisclosure in posts where the other users proactively mention the protected user, but no significant change when the other users are reacting to the protected user's posts. In addition, peer-disclosure through explicit specification, such as the direct mentioning of a user's location, decreases sharply, but no significant change occurs for implicit inference, such as the disclosure of birthday through the date of a "happy birthday" message. The design implication here is that online social networks should provide support alerting users of potential peer-disclosure through implicit inference, especially when a user is reacting to the activities of a user in the protected mode.