Literacy Demands Placed on Students in Student–GenAI Entanglements in Higher Education

Abstract Generative Artificial Intelligence (GenAI) has fundamentally changed how students engage with technology in higher education, placing them in an active epistemic role — prompting, evaluating, and co-producing with the system. Yet existing systematic reviews of AI in higher education classify AI by institutional function and were conducted largely before GenAI entered educational practice at scale, leaving a fundamental question unaddressed: what distinct forms of student–GenAI engagement exist, which are prevalent, and which remain marginal ? This paper argues that the uneven distribution of student–GenAI engagement reflects the literacy demands that different forms of engagement place on students and the degree to which those demands align with the practice logic of higher education. Drawing on Strobel et al.’s taxonomy of GenAI applications and Dohn’s context-level framework of competence demands, we classify 102 empirical studies of GenAI in higher education by entanglement type and examine the literacy demands each type generates. We find that Assistant and Enabler entanglements dominate the corpus, while Generator , Reimaginator , and Synthesizer entanglements remain marginal — a pattern that reflects not technological limitation but increasing tension between epistemic work distribution and institutional competence expectations. These findings suggest three design principles for teachers seeking to develop students’ literacy across the full range of entanglements that a postdigital world requires.