AI humanizers and the political economy of academic integrity in higher education

Abstract Artificial Intelligence (AI) has disrupted assessment in higher education and accelerated an existing cycle of performativity. Institutional policies demand the demonstration of independent authorship in assessed work, yet AI models enable students to mimic independent work that is ultimately not their own. Some institutions and educators have responded by using AI text detection technologies, although their use appears contested and uneven. Within this environment, AI humanizers have emerged. These internet-based services alter AI-generated text to evade human and technological detection. Although humanizers have attracted increasing public attention, there is a dearth of scholarly literature on the subject. This paper presents the first exploratory, systematic investigation of AI humanizer websites, framed through Goffman’s sociological account of dramaturgy. Using a systematic search and custom rubric, we catalogue 55 humanizer sites, assessing their performance of identity and conducting an in-depth multimodal critical discourse analysis (MCDA) of a purposive sample of three sites. Findings reveal that humanizers appeal to students as consumers through multiple strategies, performing legitimacy work by recontextualizing their service as a rational and defensible response to surveillance, engaging in mystification by presenting their products as proprietary, advanced technologies, and drawing on symbols of authority and trustworthiness (such as university logos) to build user confidence. We argue that humanizer services form a node in an ongoing feedback loop, in which assessment requires performances of legitimacy, and those performances create demand for technologies that enable them.