Teaching Pragmatic Appropriateness in EFL Using AI-meditated Contextual refinement

Abstract

While AI tools are increasingly used in EFL, most current applications focus on grammatical correction and lexical simplification, with limited attention to pragmatic appropriateness. this paper explores to seek the efficacy of prompt engineered LLMs in augmenting the sociopragmatics abilities of EFL learners, with a specific focus on the speech act of requesting. A quantitative survey design was adopted for this research. A total of 24 Omani undergraduate students enrolled in English language major participated. Participants evaluated AI generated requests across three different social contexts using a 3-point likert scale based on clarity, politeness and confidence in use. The AI prompt were designed based on Brown and Levinson's politeness theory, taking into consideration (social distance), (relative power) and ( ranking of imposition). participants shows preference for AI meditated sentences in academic and professional scenarios, though it faces limitations in casual scenario. This research argues that LLMs have the potential to function as adaptive tools for EFL learners, but also emphasize for need for learner autonomy to prevent "over correction" in low social distance situations.