Culturally Responsive AI Scaffolding in Collaborative Language Education: Exploring Sociocultural Adaptations across Diverse Learning Contexts

Abstract

The paper aims to review the application of culturally responsive artificial intelligence (AI) scaffolding in language learning and explore how this technology can be adapted to various sociocultural and collaborative learning environments. The key research questions include reviewing how AI tools can be integrated into culturally responsive pedagogy, identifying the most successful practices that support equal participation among all learners, and considering the restrictions and ethical issues associated with their use. It was conducted using a systematic literature review (SLR) approach to ensure a transparent and rigorous methodology. The conducted study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) to select and synthesise the relevant studies. Large databases were used to retrieve twelve peer-reviewed articles from 2020-2024. On screening and evaluation against rigorous inclusion and exclusion criteria, the studies were thematically coded and quality-appraised with the CASP checklist. The results suggest that although the AI scaffolding holds excellent promise as a means of assisting with cognitive, metacognitive, and affective aspects of language acquisition, its cultural flexibility leaves much to be desired. Few studies have explicitly dealt with cultural responsiveness, and the majority of the tools were developed based on standardised or Western-centric frameworks. Participatory design methods, the incorporation of culturally inclusive case studies, and the application of AI to facilitate intercultural dialogue are promising approaches. The findings suggest that more comprehensive AI systems should be trained on multicultural datasets and developed with the input of teachers and students from diverse backgrounds. Ethical issues, including the reduction of bias and fair accessibility to AI technologies in various socio-economic and geographical settings, are also highlighted in the review. The culturally responsive AI scaffolding should be context-aware, inclusive, and co-designed with users to ensure its effectiveness and efficacy. Longitudinal effects and the expanded representation of the Global South should be studied in the future, and policy should be informed by the equitable introduction of AI in education.