The rapid integration of artificial intelligence into higher education has reshaped how postgraduate students approach academic reading and learning. This study explores postgraduate English Education students’ perceptions of ChatGPT as a substitute for traditional reading, focusing on how the tool influences their cognitive engagement and reading behaviour. Anchored in the Technology Acceptance Model (TAM), Cognitive Offloading Theory, and Reading Literacy Theory, the research employs a quantitative survey design involving 108 master’s students from the Graduate Program of Universitas Negeri Makassar. A structured questionnaire measured four constructs: Perceived Usefulness (PU), Perceived Ease of Use (PEOU), Cognitive Offloading (COG), and Reading Habits Impact (RHI). Data were analysed using descriptive statistics, correlation, and reliability testing. The findings reveal that students perceive ChatGPT as highly useful (M = 4.23) and easy to use (M = 4.35). These positive perceptions correlate strongly with cognitive offloading (r = 0.61), indicating frequent reliance on ChatGPT to simplify learning tasks. However, a significant negative correlation between cognitive offloading and reading habits (r = –0.54) suggests that increased dependence on ChatGPT reduces students’ motivation for traditional reading. Overall, the study highlights a dual outcome, while ChatGPT enhances learning efficiency and accessibility, it simultaneously contributes to the decline of deep reading practices. The results underscore the need for balanced AI integration that promotes critical reading, reflective thinking, and responsible technology use in postgraduate education.