Understanding Unauthorized Access using Fine-Grained Human-Computer Interaction Data

Authors

Michael D. Byrd, University of ArizonaFollow
Jeffrey L. Jenkins, Brigham Young UniversityFollow
David Kim, University of ArizonaFollow
Manasvi Kumar, University of ArizonaFollow
Andrew W. Schwartz, Brigham Young UniversityFollow
Joe Valacich, University of ArizonaFollow
Parker A. Williams, University of ArizonaFollow
Ryan T. Wright, University of VirginiaFollow

Abstract

Unauthorized Data Access (UDA) by an internal employee is a major threat to an organization. Regardless of whether the individuals engaged in UDA with malicious intent or not, real-time identification of UDA events and anomalous behaviors is extremely difficult. For example, various artificial intelligence methods for detecting insider threat UDA have become readily available; while useful, such methods rely on post hoc analysis of the past (e.g., unsupervised learning algorithms on access logs). This research-in-progress note reports on if the analysis of Human-Computer Interaction (HCI) behaviors, which have been empirically validated in various studies to reveal hidden cognitive state, can be utilized as a method to detect UDAs. To examine this, an experimental design was required that would grant the subjects an opportunity to engage in UDA events while tracking the HCI behaviors in an unobtrusive manner. Background, experimental design, study execution, preliminary results, and future research plans are presented.

Recommended Citation

Byrd, Michael D.; Jenkins, Jeffrey L.; Kim, David; Kumar, Manasvi; Schwartz, Andrew W.; Valacich, Joe; Williams, Parker A.; and Wright, Ryan T., "Understanding Unauthorized Access using Fine-Grained Human-Computer Interaction Data" (2019). SIGHCI 2019 Proceedings. 13.
https://aisel.aisnet.org/sighci2019/13