Integrating AI in Pakistani ESL classrooms: Teachers' practices, perspectives, and impact on student performance

The global rise of Artificial Intelligence (AI) in English as a Second Language (ESL) education has shown promise, yet its application in resource-constrained contexts like Pakistan remains underexplored. This study examines the integration of AI tools in Pakistani ESL classrooms, with a focus on (1) teachers' instructional practices, (2) student learning outcomes, and (3) implementation challenges. Using a mixed-methods approach, data were collected through classroom observations, focus group discussions, and pre- and post-tests on vocabulary and writing skills administered to 100 undergraduate students (50 in the experimental group and 50 in the control group) over 16 weeks. The experimental group, taught with AI tools such as Grammarly and QuillBot, demonstrated significantly greater gains in vocabulary (+45%, d = 1.12) and writing performance (+46%, d = 1.03) compared to the control group. Qualitative findings revealed that while AI tools supported grammar correction and vocabulary enhancement, their effectiveness was limited by infrastructural constraints, insufficient teacher training, and cultural misalignment in language feedback. The study concludes that AI can meaningfully enhance ESL instruction when paired with teacher facilitation and localized design. It offers novel insights into culturally responsive AI integration in under-resourced educational contexts.