ChatGPT and Co-Constructed Learning: Postgraduate Perspectives on Using Artificial Intelligence in Higher Education at a Third-Culture University

Artificial intelligence (AI) is reshaping teaching and learning in higher education (HE), yet its educational meaning is context-dependent. This article reports part of a study of Chinese postgraduate perspectives on generative AI (GenAI), with a specific focus on ChatGPT as a learning technology in an English Medium Instruction (EMI) transnational setting. A mixed-methods study generated insight from 103 postgraduate students at a UK joint-venture university in China, indicating that students positioned ChatGPT as a learning assistant rather than a substitute for learning. Meanwhile, they expressed acute awareness of ChatGPT’s academic reliability and accuracy, emphasising the need for verification practices and boundaries between legitimate learning support and inappropriate substitution. Interpreted through the lens of Third Culture Universities (TCUs), an idea describing complex and evolving international campuses, the findings illustrate how GenAI becomes deeply embedded within a learning ecology shaped by imported academic norms, EMI demands, technology, and transnational quality assurance. Rather than advocating prohibition, the study highlights the need for curriculum-integrated guidance, staff development, and explicit AI literacy teaching. It concludes by proposing a curriculum-centred approach that foregrounds responsible use, transparency, and verification as core learning outcomes in the GenAI era.