AI-Powered Digital Literacy Curriculum for Multilingual Communities

Thesis Abstract

The rapid expansion of digital technologies has intensified the digital divide in multilingual communities, where heterogeneous language practices and limited access to culturally relevant ICT resources hamper equitable digital literacy development. This study addresses the problem of insufficiently inclusive digital literacy curricula that fail to accommodate linguistic diversity, thereby constraining learners’ ability to participate meaningfully in educational and civic activities in the digital age. The aim is to design and evaluate an AI-powered digital literacy curriculum tailored to multilingual learners, integrating language-aware natural language processing, culturally responsive pedagogy, and adaptive feedback mechanisms to enhance proficiency, confidence, and equitable access. Specific objectives are (1) to examine current digital literacy curricula for multilingual applicability and identify gaps; (2) to develop an AI-enabled curriculum framework that personalizes instruction across language groups using multilingual ontologies and performance analytics; (3) to implement a pilot in two urban schools with diverse language profiles and engage 420 students across grades 6–8; (4) to assess learning gains, engagement, and linguistic inclusion through mixed-methods analysis; and (5) to propose scalable, policy-relevant recommendations for integrating AI-driven digital literacy in multilingual settings. The study employs a mixed-methods design underpinned by sociocultural and instantiated learning theories, drawing on Vygotsky’s social constructivism to justify collaborative AI-supported learning and Cummins’ Linguistic Interdependence Hypothesis to frame translanguaging in digital tasks. Data were collected from a stratified sample of 420 students (approximately 60% multilingual language learners and 40% proficient L1/L2 users) across two secondary urban schools, with teacher participants (n=20) and ICT coordinators (n=6). Instruments include a standardized digital literacy assessment aligned with the curriculum’s competencies, a 5-point Likert-scale engagement survey, classroom observation protocols, and semi-structured interviews with students and teachers. Instrument validity is established through expert review and pilot testing (Cronbach’s alpha > 0. Eighty-five hours of classroom video data are coded for instructional quality and AI assistance usage using a validated coding scheme. Quantitative data are analysed using hierarchical linear modeling to examine student-level gains in digital literacy scores while accounting for nested data (students within classrooms) and language group effects; regression analyses identify predictors of achievement, including AI-assisted feedback frequency, prior digital literacy, and language proficiency. Descriptive statistics and thematic analysis of interview transcripts and observational notes triangulate findings on learner engagement, perceived inclusivity, and instructional appropriateness. The study also employs ANOVA to compare outcomes across language groups and schools, with post-hoc tests to identify specific differences. Model specification incorporates latent growth curves to capture trajectory trends in digital literacy competencies over the intervention period. Anticipated findings indicate that the AI-powered curriculum yields statistically significant improvements in digital literacy proficiency for multilingual learners relative to baseline and to the control condition, with larger effect sizes observed in tasks requiring code-switching, translation-aware problem solving, and collaborative annotation. Engagement is expected to be higher in AI-adaptive tasks that provide immediate, culturally contextualized feedback, and teachers report increased instructional efficiency due to AI-enabled dashboards for monitoring learner progress. The study also anticipates evidence of enhanced translanguaging practices and more equitable participation across language groups, mediated by the integrated multilingual ontologies and feedback loops embedded in the curriculum. The contribution to knowledge includes (i) a novel AI-enabled, linguistically inclusive digital literacy curriculum framework grounded in sociocultural theory and translanguaging perspectives; (ii) empirical evidence on the efficacy of language-aware AI feedback and adaptive instruction for multilingual learners; (iii) a scalable implementation model with clear guidelines for teacher professional development, assessment alignment, and ICT infrastructure requirements; and (iv) policy-oriented recommendations for national or regional curricula to embrace AI-assisted digital literacy within multilingual education ecosystems. The study concludes that AI-driven, language-sensitive digital literacy curricula can significantly reduce the digital divide in multilingual communities when designed with robust pedagogical grounding, rigorous evaluation, and transparent, ethical AI practices. Recommendations include expanding pilot programs to additional language ecosystems, investing in bilingual AI training data and interpretability tools for educators, and establishing governance frameworks to safeguard data privacy and cultural relevance in AI-assisted learning environments.