Design and evaluate a chatbot for language learning in multilingual contexts

Thesis Abstract

In an increasingly interconnected global landscape, multilingual communication is essential, yet language learning in diverse linguistic environments remains a significant challenge due to limited access to effective, contextually relevant pedagogical tools. This study aims to design, implement, and evaluate a chatbot tailored to facilitate language acquisition within multilingual settings, with a focus on enhancing interactive learning, learner engagement, and linguistic competence. The specific objectives include developing a chatbot that supports multiple languages, integrating speech recognition and natural language processing (NLP) features, assessing its impact on language proficiency, and determining user satisfaction and engagement levels. The research adopts a mixed-methods design, combining quantitative experimental assessment with qualitative insights to comprehensively evaluate the chatbot’s effectiveness. The target population comprises adult language learners enrolled in university language programs across multiple institutions in the country, totaling an estimated 500 participants. A stratified random sampling technique selects a sample of 200 learners, ensuring diverse representation across linguistic backgrounds, proficiency levels, and age groups. Data collection instruments include pre- and post-intervention language proficiency tests, structured questionnaires measuring learner attitudes, engagement metrics generated by the chatbot, focus group discussions, and semi-structured interviews. The language tests are scored using standardized assessment criteria aligned with the Common European Framework of Reference for Languages (CEFR), while questionnaires and interviews employ validated scales for measuring learner motivation and satisfaction. The analysis employs statistical techniques such as paired t-tests and ANOVA to compare pre- and post-intervention proficiency scores, along with regression analysis to identify predictors of language improvement. Thematic analysis is applied to qualitative data from learner feedback and interview transcripts, identifying common themes related to user experience and perceived efficacy. The chatbot’s development process follows an iterative design framework, incorporating user-centered design principles, with implementation guided by established theories such as Vygotsky’s Social Constructivism and the Cognitive Load Theory, to foster meaningful interaction and optimize the learner’s cognitive engagement. Expected findings suggest that the chatbot significantly enhances language learning outcomes, evidenced by measurable increases in proficiency scores and positive shifts in learner attitudes towards language practice. It is anticipated that learners will report higher engagement levels and perceive the chatbot as a convenient, supportive tool for autonomous learning. The study also expects to identify specific features—such as real-time feedback and multilingual support—that positively influence learner motivation and satisfaction. This research contributes to the existing body of knowledge by providing an empirical evaluation of chatbot-based language learning tools in multilingual contexts, highlighting their potential to complement traditional pedagogies and bridge resource gaps. It offers practical insights into the design and deployment of AI-driven language learning aids, grounded in contemporary learning theories and technological best practices. The findings are relevant for educators, language instructors, and developers seeking to harness artificial intelligence to improve multilingual language acquisition. The study concludes that well-designed chatbots can serve as effective, scalable learning aides that support language learners in diverse settings, fostering autonomous and sustained engagement. Recommendations include adopting user-centered design strategies, integrating speech recognition capabilities for pronunciation practice, and continuous iterative improvements based on learner feedback. Future research should explore longitudinal effects, expand to different language pairs, and investigate pedagogical integration within formal curricula to maximize the technological potential for multilingual language education.