Privacy: Gone with the Typing! Identifying Web Users by Their Typing Patterns

Abstract

The lack of privacy protection for Internet users has been identified as a major problem in modern web browsers. Despite potentially high risk of identification by typing patterns, this topic has received little attention in both the research and general community. In this paper we present a simple but efficient statistical detection model for constructing users' identity from their typing patterns. Extensive experiments are conducted to justify the accuracy of our model. Using this model, online adversaries could uncover the identity of Web users even if they are using anonymizing services. Our goal is to raise awareness of this privacy risk to general Internet users and encourage countermeasures in future implementations of anonymous browsing techniques.

Exploring Touch-Screen Biometrics for User Identification on Smart Phones

“…Related to the work reported in [5], our results imply that, regardless of the users' choices of patterns for authenticating to different service providers, attackers would have a bigger probability to uncover the identity of pseudonymous users based on the biometrics of their secret lock patterns drawn at different websites, assuming that a website can get a hold of users' biometric data. Therefore, using lock patterns as a remote authentication method could result in linkability attacks and user profiling even when using different pseudonyms and patterns as passwords at different services' sites.…”

“…Although these events allow users to browse the Internet in a more user-friendly manner by dragging their finger around to pan a webpage or elements contained within a webpage, this implies that, by implementing these events, any common website 4 U-PrIM (Usable Privacy-enhancing Identity Management for smart applications) http://www.kau.se/en/computer-science/research/research-projects/u-prim 5 An example of the touch and drag events for Firefox on Android devices can be found at http://www.quirksmode.org/m/tests/drag.html (Accessed: 2011-10-27).…”

“…This probability of linking and profiling users could increase if users' pseudonyms can be identified based on their lock pattern behavior when authenticating into different services. Yet another study has shown that users' identities can be reconstructed from their typing patterns while browsing online [5]. In our paper, we consider if similar issues can arise when employing the biometrics of lock patterns for authentication.…”

CADIVa: cooperative and adaptive decentralized identity validation model for social networks

“…[Cai et al, 2011] suggests people to people recommendations for friendships' acceptance by relying on collaborative filtering techniques. In [Chairunnanda et al, 2011], users are suggested to be identified from their typing patterns; whereas chatting patterns are exploited for users' identification in [Roffo et al, 2013]. More recently, [Goga et al, 2013] suggests identifying users across networks based on geo-location and time-stamp information attached to their posts and on their writing styles.…”

“…For example, some of them identify users by utilizing their sensitive information such as geo-locations they usually visit and time-stamps of the information they share [Goga et al, 2013]. In [Chairunnanda et al, 2011], authors use typing patterns to identify users, whereas chatting patterns are exploited in [Roffo et al, 2013]. Additionally, other validation approaches have suggested to rely on human feedback.…”