The Active Learning–GenAI Synergy Framework: Ethical Integration of Generative AI in EFL/ESL Writing in Resource-Constrained Contexts

TL;DR

This study explores how generative AI tools can support active learning in EFL writing in resource-limited settings, showing improved student performance and positive attitudes.

Contribution

A practical model for ethical AI integration in EFL writing that emphasizes pedagogical structures over technology alone.

Findings

Students using AI tools showed significant writing performance improvements compared to traditional instruction.

AI tools were perceived as cognitive partners, metacognitive mirrors, and equity tools by most participants.

Positive attitudes toward AI use were reported by 79.7% of students, with 86% intending to continue using AI.

Abstract

This mixed-methods action research study investigates the integration of ethically designed generative AI tools as catalysts for active learning in English as Foreign Language (EFL) writing instruction. The research is situated within resource-constrained higher education contexts and is grounded in Kolb’s experiential learning theory and active learning principles.

The study employed a quasi-experimental design involving 148 undergraduate students from Bahauddin Zakariya University, Pakistan. Over a 15-week intervention, an experimental group utilized AI tools (ChatGPT, Claude AI, Meta AI, and Canva) within a cognitive partnership model, while a control group received traditional teacher-centered instruction. Quantitative data on writing performance was supplemented by qualitative data from semi-structured interviews to capture student experiences and the development of ethical awareness.

Quantitative results show the experimental group achieved statistically significant improvements in writing performance (Z = -6.325, p < .001) compared to modest gains in the control group (Z = -2.128, p = 0.033), with notable skill progression emerging after 6-8 weeks. Qualitative analysis revealed that AI tools successfully functioned as cognitive partners, metacognitive mirrors, and equity tools. A strong majority of participants (79.7%) expressed positive views about AI integration, with 86% indicating intentions for continued use.

The paper provides a practical model for embedding AI tools in collaborative-learning settings while upholding academic integrity through a disciplined process of ethical reflection. The findings challenge the presumption that technology alone drives educational improvement, demonstrating instead that it is the pedagogical structures mediating AI interaction that determine an effective and ethical educational process.