A study on the mathematical learning knowledge structures of junior high school mathematically gifted students in Nanning, China.

Generative artificial intelligence has become increasingly visible in higher education, where students use AI tools for information search, writing support, translation, coding assistance, and problem solving. Yet its promotion and practice at local undergraduate universities in five northwestern Chinese provinces differ sharply from resource-rich universities in eastern China. Restricted by regional fiscal investment, digital infrastructure and students’ baseline digital literacy, these universities face inadequate resource supply. Such shortages emerge in unstable campus networks, insufficient intelligent teaching facilities, limited access to paid AI tools, and scarce training on AI operation. Existing studies have begun to examine regional disparities in Chinese college students’ adoption of generative AI (). Still, most research relies on convenient samples from central cities and well-resourced universities. Few empirical studies specifically target local undergraduate schools in Northwest China where educational resources fall short. Against this backdrop, students’ sustained use of AI educational tools depends on more than perceived ease of use. It also hinges on their ability to verify AI-generated content, refine prompts, correct flawed outputs, and maintain consistent AI learning amid insufficient institutional policies and training support. Moreover, educational applications of generative AI may widen the new digital divide. coined the term “GenAI divide” among college students. This divide covers not only access to AI tools but also AI literacy, usage confidence, institutional support, and the capacity to translate AI into academic advantages. Thus, insufficient resource supply is not merely a problem of tool accessibility. It may also hinder students’ development of effective, sustained, high-quality AI-assisted learning practices. This study targets local undergraduate universities across five northwestern provinces. It explores the formative mechanisms underlying college students’ sustained intention to use AI educational tools.