Natural Language Processing (NLP) is a science aimed at computationally interpreting written language. This field is maturing at an extraordinary pace. It is creating significant value and advancing a number of key research fronts. However, it also enables highly sophisticated phishing attacks. Given a large enough text sample, an NLP algorithm can identify and replicate defining characteristics of an individual’s communication patterns. This facilitates programmatic impersonation of trusted individuals. A natural language processor could interpret incoming text messages or email and improvise responses which approximate the language of a known contact. The recipient could be tricked into sharing sensitive information. Just how vulnerable are we? This paper reviews the state of the art of natural language processing and social engineering. It also describes a test which empirically assesses our ability to discern legitimate communications from algorithmically-produced forgeries.
Shropshire, Jordan, "Natural Language Processing as a Weapon" (2018). WISP 2018 Proceedings. 26.