Fighting fraud: Corpus-assisted approaches to understanding and disrupting fraud activity on the dark web

Financial fraud has risen steeply over the last decade and, according to data from the National Crime Agency, is currently recognised as the most commonly experienced crime in the UK, accounting for over 40% of all crimes in England and Wales committed against individuals over 16. Much of this increase is attributed to the rise and evolution of online technologies which have ushered in a wave of new methods and opportunities for perpetrators as well as an era of unprecedented personal self-disclosure via social media by potential victims whose details can be readily exploited.
A key affordance to perpetrators is the rise of illicit marketplaces and crime-focused discussion fora on the dark web, i.e. a portion of the internet unindexed by mainstream search engines. Such spaces provide users a level of anonymity that makes policing them very difficult, yet they are fruitful sites for linguistic exploration regarding the behaviours and activities of the relevant communities of practice. We demonstrate the application of corpus methods to addressing online fraud by, firstly, showing how a linguistically-informed understanding of online fraud communities’ interactions can assist the undercover policing of dark-web fraud fora with regard to the specific task of community infiltration. Secondly, we address the problem from a commercial perspective, demonstrating how corpus analytic methods can inform online tools designed to help commercial entities monitor dark-web spaces for fraud activity related to their products, and how popular corpus tools can be tweaked for use by non-linguist audiences for this purpose.