AI AS ASSISTANT OR SUBSTITUTE? NEGOTIATING LEARNER AUTONOMY AND TECHNOLOGICAL DEPENDENCE IN AI-MEDIATED ENGLISH LANGUAGE LEARNING

Abstract

Generative artificial intelligence (AI) is increasingly becoming part of how students learn English, offering immediate explanations and opportunities to practice beyond the classroom. Yet, the convenience of these tools raises a growing concern. This study explores how university-level English as a Second Language (ESL) learners negotiate this boundary in their everyday use of generative AI for English language learning. Guided by learner autonomy and Oxford's Strategic Self-Regulation (S2R) Model, the study views learners as active decision-makers who plan, monitor, evaluate, and regulate their learning while interacting with AI. Although recent studies have examined students' acceptance, perceptions, and learning outcomes associated with generative AI, less is known about how learners experience the tension between technological assistance and dependence, particularly in Pakistani higher education. A qualitative multiple-case study will therefore be conducted with 25 university ESL learners purposively selected from higher education institutions in Pakistan. Data is generated through semi-structured interviews and examined using reflexive thematic analysis. The study shows that learners move between using AI as a source of scaffolding. The study seeks to extend current understanding of AI-mediated language learning by highlighting how learner autonomy can be both supported and challenged by the same technology, with implications for more critical and responsible use of AI in English language education.