The Nexus of Gen‐ AI Features in Higher Education: Reducing Cognitive Load, Building Resilience, and Personalising Learning

ABSTRACT Background Generative Artificial Intelligence (Gen‐AI) is reshaping higher education by offering adaptive support, instant feedback, and greater learner autonomy. Despite the increasing use of tools such as ChatGPT in university learning, empirical evidence on how their motivational affordances influence academic resilience and personalised learning remains limited, especially among non‐native learners. Objectives This study examines how five motivational features of Gen‐AI‐Adaptive Learning Support (ALS), Instantaneous Problem Resolution (IPR), Subject‐Specific Guidance (SSG), Temporal Flexibility (TF), and Language Accessibility (LA) influence cognitive load, academic resilience, and personalised learning. Specifically, it addresses three research questions: (1) To what extent do generative AI affordances influence cognitive load reduction among higher education students? (2) To what extent do generative AI affordances influence academic resilience and personalised learning? (3) Does cognitive load reduction mediate the relationship between generative AI affordances and academic resilience and personalised learning? Methodology A cross‐sectional survey was conducted with 813 non‐native university students from seven higher education institutions in China. Structural equation modelling was used to test the hypothesised relationships among Gen‐AI affordances, cognitive load, academic resilience, and personalised learning. Results All five motivational features positively influenced cognitive load reduction, personalised learning, and academic resilience. Cognitive load reduction significantly mediated these relationships, suggesting that cognitive load reduction represents an important indirect pathway linking Gen‐AI affordances with learning efficiency, academic resilience, and personalised learning. Conclusion Gen‐AI can support personalised and resilient learning by reducing cognitive burden. This study clarifies the mediating role of cognitive processes and offers practical implications for educators and EdTech developers designing adaptive and inclusive learning environments.