Adaptive Mobile Micro-Learning for Adult Language Learners via AI Tutor

Thesis Abstract

This study investigates how adaptive mobile micro-learning, delivered via an AI-powered tutor, can enhance language acquisition among adult learners who balance work, family, and study commitments in urban settings. The problem addressed is the inefficiency of traditional, time-intensive language programs for adults, which often fail to sustain daily engagement and fail to personalize content to individual cognitions, prior knowledge, motivation, and contextual language needs. The aim is to develop and evaluate an adaptive mobile micro-learning system that personalizes short, focused language activities grounded in evidence-based adult learning theory to improve communicative proficiency, retention, and autonomous learning skills. Specific objectives include (1) designing an AI tutor that dynamically sequences micro-learning episodes (10–15 minutes each) based on learner profiles, affective state, and performance; (2) implementing natural language processing (NLP) and reinforcement learning components to tailor content, feedback, and difficulty; (3) assessing the impact of adaptive micro-learning on receptive and productive language skills using standardized tests; (4) examining changes in learner engagement, self-regulated learning strategies, and time-on-task; and (5) exploring learners’ perceived usefulness and satisfaction with the mobile intervention. The study employs a mixed-methods design combining a quasi-experimental framework with qualitative inquiry. A sample of 240 adult language learners, aged 25–50, recruited from three metropolitan language centers and corresponding online communities, will be randomly assigned to an experimental group (adaptive AI-tutor micro-learning) and a control group (non-adaptive mobile micro-learning with static content). Data collection will involve pre- and post-tests of language proficiency (Pearson CEFR-aligned assessments, including listening, speaking, reading, and writing components), weekly engagement analytics (time-on-platform, lesson completion rate, and retention), and validated self-report instruments measuring motivation (Motivational Regulation Scale for L2 learning) and self-regulated learning strategies (MSLQ). The AI tutor will log interaction data across 12 weeks of intervention. Instrument reliability will be established via Cronbach’s alpha (aiming for ? ? 0.80) and test-retest stability. Content validity will be confirmed by a panel of language pedagogy and AI-education experts. Quantitative analyses will include ANCOVA to compare post-test outcomes between groups while controlling for baseline proficiency, multilevel modeling to account for repeated measures and nested time points, and regression analyses to identify predictors of language gains and engagement. Mediation analyses will test whether engagement and self-regulated learning mediate the relationship between adaptive tutoring and proficiency improvements. Qualitative data will be gathered through semi-structured interviews with 24 participants (12 from each group) and two focus groups with tutors, analyzed using thematic analysis to identify perceived benefits, challenges, and suggestions for system refinement. The study will triangulate quantitative outcomes with qualitative insights to elucidate mechanisms of change. Expected findings include significantly greater gains in speaking and listening proficiency for the experimental group, higher engagement metrics, and improved self-regulated learning strategies. The AI-driven personalization is anticipated to demonstrate superior content alignment with learners’ ZPD (Zone of Proximal Development) and affective states, yielding more efficient practice cycles and higher satisfaction. Theoretically, the study extends socio-constructivist and self-determination theories by demonstrating how mobile, micro-learning environments scaffold collaborative and autonomous learning through adaptive feedback loops. It also contributes methodologically by integrating reinforcement learning-based content sequencing with rigorous evaluation in adult language contexts. The study’s contribution to knowledge lies in (i) providing empirical evidence on the efficacy of adaptive mobile micro-learning for adult language development, (ii) detailing a scalable, AI-enabled framework for personalized language practice, and (iii) offering design guidelines for future AI tutors in adult education that address time constraints, motivation, and retention. Practical implications include informing the development of commercially viable, evidence-based language learning apps and informing policy on the integration of AI-enabled micro-learning into continuing education programs. Overall, the study concludes that adaptive mobile micro-learning via AI tutors can significantly enhance language outcomes for adults while promoting sustained engagement and self-regulated learning, recommending further research into multilingual support, long-term retention, and cross-cultural applicability.