Keystroke Dynamics and Identifying Cyber-Attackers

The growth of commerce on the Internet has been accompanied by increasing online crime. Cyber security was identified as a serious national security problem by President Obama and national security officials in 2011. Identifying perpetrators of online crimes or attacks is essential not only for law enforcement, but also from a national security perspective. The inability to catch and prosecute online criminals has been due to the difficulty in identifying and attributing a crime to any individual. In one promising line of research, a person's online typing habits and characteristics have been used to distinguish one person from another based solely on the features of a person's keystrokes. The acceptance of keystroke patterns as a tool for security professionals and law enforcement to identify online criminals is an area that has not been explored and is key to whether this newer biometric will be accepted for identification similar to physical fingerprints. Once the identification of a suspect is made then those interested in prosecution can focus their efforts on that person and use other forms of evidence to support prosecution if needed. Determining whether keystroke dynamics is admissible as evidence was outside the scope of this study. The specific problem addressed is the inability to identify online criminals and the acceptance of keystroke dynamics as a tool to help track and identify those criminals. The purpose of this quantitative correlational study was to identify which factors predict the intention to use keystroke patterns as a means to identify online criminals. A survey of security professionals and law enforcement investigators on the acceptance of such a tool and their intentions was conducted. The participants included IT professionals and law enforcement professionals who work with computer forensics. Over 500 participants from two organizations were invited to participate which yielded 199 valid responses. The results indicated that all five of the predictor variables had statistically significant bivariate correlations with the outcome variable of behavioral intention to use keystroke patters as an investigative tool. Future research could benefit from this and build on the findings to study what characteristics of keystroke dynamics are unique and most beneficial in creating a digital fingerprint of individuals typing habits.