Developing an “Insider Language Index” as a Composite Measure to Detect Insider Threat

Abstract

Of all the issues that confront modern organisations, insider threat is one of the most challenging in terms of impact and mitigation. Typically, research has focused on defining what insider threat is and determining how such threats can be detected through technology. We suggest that technological approaches to insider-threat detection can be complemented through a greater focus on investigating the linguistic behaviours associated with insider activity. Research has highlighted that an individual’s use of language offers a potential means of identifying insiders. Using Linguistic Inquiry and Word Count, this study analysed the language used by insiders and non-insiders during workplace interviews. Results revealed that, compared to non-insiders, insiders used significantly more words relating to cognitive processing, significantly more self-referential terms, and significantly more negative emotion words. Based on these findings, a generalisable Insider Language Index (ILI) was developed that has the potential to support insider detection in organisational contexts.