Speaker
Description
This study explores the design and implementation of a "closed-loop GenAI ecosystem" to address the severe lack of authentic language input and speaking opportunities in rural primary English Language Teaching (ELT), while preparing young learners for high-stakes online competitions.
Utilizing an action research design over a two-year period, this initiative integrated three distinct Generative AI tools into a synergistic learning loop for Grade 3 students in a resource-constrained Vietnamese primary school. The ecosystem operated in three synchronized phases:
Input and Phonics Synthesis (Suno AI): Converting textbook vocabulary and grammatical patterns into catchy, customized educational songs to enhance phonological awareness and vocabulary retention.
Contextual Knowledge & Resource Curation (NotebookLM): Synthesizing mock test materials and generating personalized study guides tailored to individual learning paces.
Real-time Conversational Practice (Gemini Live): Serving as an on-demand, interactive speaking partner to build confidence and oral fluency.
The pedagogical intervention yielded exceptional quantitative and qualitative outcomes. Within two years of implementation, this cost-effective model successfully bridged the local digital divide, culminating in historic academic achievements: 14 students qualified for the provincial-level Internet Olympiads of English (IOE), and 7 secured national-level prizes—marking the first national-level academic accolades in the school’s history. Qualitatively, the ecosystem democratized high-quality language practice, sparking a self-sustaining English learning movement within the local community.
This paper provides a scalable, pedagogically grounded blueprint for language educators seeking to harness free-access GenAI tools to cultivate communicative competence and elite competitive readiness in marginalized educational settings.
Keywords: Closed-loop GenAI, NotebookLM, Suno AI, Gemini Live, Primary ELT, Internet Olympiads of English (IOE).
