Empowerment or Divide? A Mixed-Methods Investigation into AI Tool Usage in Chinese Language Learning among International Students in China’s Higher Vocational Colleges

Abstract

The proliferation of generative artificial intelligence (AI) tools—such as large language models, machine translation applications, and intelligent writing assistants—has profoundly transformed language learning practices worldwide. For international students in China, these tools present unprecedented opportunities to overcome linguistic barriers and enhance learning efficiency. However, concerns are growing that differential access, digital literacy, and usage patterns may inadvertently generate a new form of “digital divide” within this already vulnerable population. Drawing upon van Dijk’s (2024) multidimensional digital divide framework and the Technology Acceptance Model (TAM), this study employs a sequential explanatory mixed-methods design to investigate AI tool usage among international students in three Jiangsu higher vocational colleges. In the quantitative phase, 256 international students completed a structured survey assessing AI usage patterns (types, frequency, scenarios), perceived empowerment (learning efficiency, personalized support, motivation), and digital divide indicators (technological dependency, critical thinking erosion, academic integrity concerns). Hierarchical regression and mediation analyses were conducted to identify significant predictors and underlying mechanisms. In the qualitative phase, 18 purposefully selected participants engaged in semi-structured interviews to elaborate on quantitative findings and capture lived experiences. Preliminary results indicate that while over 89% of participants regularly use AI tools for Chinese learning, usage patterns vary significantly by HSK level, digital literacy, and teacher guidance. Empowerment effects are most pronounced in vocabulary acquisition and writing assistance, whereas divide effects are most evident in higher-order thinking tasks and independent problem-solving. Teacher support and AI literacy training emerged as critical moderating factors that buffer against negative consequences. This study extends digital divide theorizing to the context of AI-enhanced language education and offers empirically grounded recommendations for fostering responsible AI use in international Chinese language education.