This comparative study critically examines the integration of artificial intelligence (AI) into early English language learning environments within the under-resourced rural contexts of Pakistan and China. Focusing on the affordances and constraints of AI-mediated tools such as automated speech recognition for phonological training and adaptive systems for personalized practice the research interrogates how these technologies are enacted within contrasting policy and sociolinguistic ecologies. Employing a mixed-methods design, it synthesizes data from documentary policy analysis, semi-structured interviews with educators and caregivers, and a review of deployed applications to construct a nuanced cross-national analysis. Findings delineate two predominant implementation paradigms: a state-coordinated, platform-centric model prevalent in China, which aligns AI supplementation with national curricular objectives, and a decentralized, NGO-facilitated model emerging in Pakistan, characterized by innovative low-resource adaptations. The analysis reveals that beyond infrastructure, critical mediating factors include the tools’ linguistic sensitivity to local dialects, their culturally embedded design, and their role in either augmenting or inadvertently displacing teacher pedagogical authority. The study contributes to Computer-Assisted Language Learning (CALL) literature by proposing a socio-technical framework for ethical AI integration in linguistically diverse, low-resource settings. It concludes with targeted recommendations for participatory, equity-driven design, underscoring the necessity of developing multilingual corpora, fostering teacher-AI collaboration, and instituting robust data governance for young learners to ensure these technologies support rather than undermine equitable language acquisition.
Ramzan et al. (Mon,) studied this question.