Between regulation and accessibility: how Chinese university students navigate global and domestic generative AI

Generative AI is rapidly reshaping higher education, yet students in China face barriers to accessing global tools such as ChatGPT due to regulatory and technological constraints. Guided by the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2), this study employs qualitative interviews to examine how Chinese university students engage with global and domestic generative AI in their learning. Findings show that accessibility and cost strongly shape adoption. Limited access to global tools and the expense of paid versions led many students to rely on domestic alternatives, though some used VPNs or shared accounts to access tools perceived as more beneficial. Language proficiency and disciplinary background also influenced engagement: STEM and English-medium students preferred global AI for technical or English-based tasks, while humanities students favoured domestic AI for Chinese-language and culturally grounded assignments. Students further viewed domestic AI as more aligned with local norms and political contexts, whereas global systems often lacked cultural nuance. By illuminating the accessibility, linguistic, disciplinary, and cultural dimensions of AI adoption, this study advances understanding of generative AI integration in non-Western higher education and calls for human-centered, multilingual, and contextually responsive approaches to foster equity in digital learning.