THE SPEAK-BOT FRAMEWORK FOR CONTEXTUALIZED ENGLISH LEARNING

Abstract

This study developed and validated the SPEAK-BOT framework, an AI-assisted model for contextualized English--speaking instruction that integrates ChatGPT interaction, teacher facilitation, and local dialogue materials from Haba Inggreh. Using the Design and Development Research (DDR) approach, the study was conducted in two high schools in Nagan Raya, Aceh—SMAN 1 Seunagan and SMAN 1 Kuala—representing agrarian and coastal contexts. A total of eighty Grade XI students and four English teachers participated in the limited implementation phase. Data were collected through expert validation, observation, questionnaires, interviews, and AI interaction logs, then analyzed using descriptive statistics and thematic interpretation. The results show that students and teachers demonstrated high motivation and readiness to use AI for speaking practice, despite minor barriers such as internet instability and limited digital familiarity. The framework successfully addressed key weaknesses in pronunciation and contextual comprehension, while enhancing learners’ confidence through enjoyable, collaborative, and story-based activities. Teachers confirmed that AI feedback complemented classroom instruction and strengthened contextual engagement. The findings imply that SPEAK-BOT effectively bridges technology, pedagogy, and local culture, showing that AI can humanize rather than mechanize English learning. The study closes the gap between AI-assisted learning and contextualized pedagogy by uniting them within a single instructional framework. Although limited in scope, the research provides a scalable foundation for broader implementation and future longitudinal studies.