Reframing and Broadening Adversarial Stylometry for Academic Integrity

Adversarial stylometry can be a form of academic fraud that involves the deceptive manipulation of text through manual, hybrid, or automatic techniques to obfuscate or impersonate with the intent to circumvent plagiarism detection; exploit or mask personal identifiers (including demographic, personality, and behavior), alter time series linguistic features, either for gain or to avoid undue discrimination. This term was reframed for academic integrity in this chapter to understand this vulnerability fully. A threat model was created that identified 29 threats classified as high, moderate, and low. Text-matching software is starting to integrate stylometry to capture one’s stylistic writing voice, particularly given the problem of contract cheating. New ways to circumvent progress come with any technical advancement, and an adversarial response is often generated. Higher education needs to be aware given its text-matching software usage. This chapter builds on the early work by Brennan, Afroz, & Greenstead (ACM Transactions on Information and System Security, 15(3), 1–22, 2011), and reviewed the literature on adversarial stylometry, emphasizing the intersection with plagiarism and contract cheating. Research examines academic integrity through the lens of forensic linguistics and looks to the field of fake news to provide new insights. This work is relevant for scholars and practitioners interested in academic integrity, text- matching software, and security.