Investigating EFL Learners’ Intercultural Pragmatic Competence and Adaptation Strategies in Interactions with Large Language Models

Le Ngoc My (Thai Nguyen College, Thai Nguyen Province, Vietnam)

International Journal of Sustainable English Language, Education, and Science
Vol. 3 No. 2 (2026)

This sequential explanatory mixed-methods study investigated the extent to which 148 Vietnamese EFL learners recognize and navigate cultural nuances, implicit biases, and pragmatic appropriateness in English responses generated by Large Language Models (LLMs). Quantitative survey data revealed a high capacity among learners to identify overt conversational directness, yet they demonstrated a moderate ability to detect deep structural erasures of non-Western scripts. Multiple linear regression analysis proved that baseline Intercultural Communicative Competence (ICC) specifically behavioral flexibility and cultural knowledge powerfully predicted 56% of the variance in learners’ AI Intercultural Literacy. Qualitative screen-recording logs and retrospective interviews with 12 typical-case participants mapped an array of adaptive prompting maneuvers, including contextual persona conditioning and socio-pragmatic constraints, used to overwrite algorithmic defaults. The findings highlight the critical necessity of integrating algorithmic critique and culturally responsive prompt engineering into contemporary university digital literacy curricula to protect human communicative agency in AI-mediated communication.