The mediating role of social influence in ChatGPT-assisted English learning performance: A sociocultural analysis in Indonesian higher education

Abstract

This study investigates the mediating role of social interaction in the relationship between sociocultural factors and English language learning performance using ChatGPT as an AI-integrated learning tool in Indonesian higher education. As the integration of generative AI like ChatGPT becomes increasingly relevant to global education, this research explores how sociocultural elements influence students’ performance through the lens of social influence. Employing a quantitative methodology, data were gathered from 220 undergraduate students across multiple universities in Jakarta through a structured online survey. Structural Equation Modeling (SEM) using LISREL 8.80 was applied to examine the relationships among constructs: sociocultural influences, social influence, and learning performance. Findings reveal a significant positive effect of ChatGPT-assisted learning on students' English proficiency, mediated by motivation (β = 0.41, p < 0.001). Social influence was found to significantly mediate the relationship between sociocultural factors and performance (indirect effect = 0.29, R² = 0.38), underscoring the importance of peer dynamics in technology-driven education. Theoretically, the study extends Vygotsky’s sociocultural learning framework and integrates the Technology Acceptance Model (TAM) and Cultural Intelligence Theory (CIT), emphasizing that cultural background and social interaction are key to the adoption and success of AI in education. Practically, the findings suggest educators and policymakers must consider social and cultural dimensions when designing curricula that integrate AI tools like ChatGPT. The research advocates for inclusive, socially engaging, and culturally sensitive digital learning environments to optimize AI adoption and enhance English language learning outcomes in multilingual contexts.