Abstract
Artificial Intelligence (AI) is rapidly transforming the educational landscape, offering unprecedented opportunities for innovation in second-language learning. At Xi'an Jiaotong-Liverpool University (XJTLU), SPA001: Spanish Language 1—an entry-level Spanish module for absolute beginners—has developed two AI-driven tools specifically designed to address students’ learning needs: AI-migo, a virtual tutor that provides tailored support across multiple aspects of learning, and XiSPA, an AI-driven conversation partner that facilitates interactive and level-appropriate communicative practice. Both tools employ a prompt-based approach that incorporates structured knowledge bases to ensure support aligns with curriculum progression.
This article presents the background, development, pilot implementation, and initial evaluation of these tools. A mixed-methods evaluation with 385 students revealed positive perceptions across ease of use, enjoyment, output quality, and perceived learning benefits, with convenience emerging as a particularly salient advantage. Students reported substantial tool adoption, with continued post-assessment use. Qualitative findings identified areas requiring refinement, including technical reliability and accessibility, interaction burden, and adaptation mechanisms, among others. The complementary roles of AI-migo and XiSPA, combined with the prompt-based approach, suggest potential for scalable, resource-efficient AI integration in beginner language education.
This article presents the background, development, pilot implementation, and initial evaluation of these tools. A mixed-methods evaluation with 385 students revealed positive perceptions across ease of use, enjoyment, output quality, and perceived learning benefits, with convenience emerging as a particularly salient advantage. Students reported substantial tool adoption, with continued post-assessment use. Qualitative findings identified areas requiring refinement, including technical reliability and accessibility, interaction burden, and adaptation mechanisms, among others. The complementary roles of AI-migo and XiSPA, combined with the prompt-based approach, suggest potential for scalable, resource-efficient AI integration in beginner language education.
| Original language | English |
|---|---|
| Media of output | Online Article |
| Publisher | Centre for Educational Innovation and Excellence, Learning Mall, XJTLU |
| Number of pages | 24 |
| Publication status | Published - 2025 |