UNCOVERING THE AI PARADOX IN UNDER-RESOURCED ESL CONTEXTS: QUANTIFYING LEARNER ADOPTION, AUTONOMY, AND DEPENDENCY

Abstract

This present study quantifies the divergent trends of the use of artificial intelligence (AI) tools, development of learner autonomy, and cognitive dependence of English as a Second Language (ESL) learners and teachers in under-resourced higher education institutions (HEIs) of South Punjab, Pakistan. This study adopted a descriptive survey design, validated instrument with 30 items on a Likert scale (Cronbach’s α = .86), and data obtained from 80 respondents comprising 60 students and 20 ESL teachers from public and private universities in Multan and Dera Ghazi Khan during the academic year 2025–2026. The results show that 83% of students and 90% of teachers are active AI users, but there is a clear Monoculture Trap present as 88% of students primarily use ChatGPT with 65% of teachers. An independent samples t-test was used to confirm the difference between autonomy and dependency as being statistically significant (d = 0.91). A regression analysis revealed that the frequency of use of AI accounted for 45% of the variance in dependency scores (R² = .45, p < .001); and there was a significant Pearson correlation between frequency of use of AI and dependency scores (r = −.75, p < .001) with a strong negative relationship. In all measures of critical thinking, students rated the role of AI in critical thinking lower than teachers, suggesting metacognition of the limitations of technology and over-reliance. This study suggests two measurable constructs, learner autonomy and AI-dependency, that are operationalized at the researcher level and can be used to ESL contexts in the Global South that are mediated by AI. The Triple-A Pedagogical Framework (Access, Audit, and Apply) is suggested to be a scalable, evidence-based model for integrating AI into the classroom without sacrificing crucial language skills. Recommendations are provided for curriculum planners, institutional administrators and policymakers in education settings where resources are limited.