Teacher perspectives on ChatGPT integrated vocabulary instruction in a flipped ESL classroom

Artificial intelligence (AI) has gained importance and popularity in education recently. Among AI tools, ChatGPT has attracted growing interest for its potential to support language learning. This study explores how ChatGPT-integrated pre-class vocabulary tasks influence student engagement in a flipped classroom among Malaysian primary school ESL learners. Specifically, it examines (i) how students engage with ChatGPT during preclass vocabulary learning, (ii) what factors influence meaningful or surface-level engagement, and (iii) how teachers perceive the challenges and opportunities of integrating ChatGPT in vocabulary instruction. This paper draws on qualitative data from teacher interviews conducted as part of a larger mixed-methods study. The interviews were analysed using Braun and Clarke’s six-phase thematic analysis approach. Findings revealed that ChatGPT initially fostered curiosity and vocabulary exploration, promoting schema activation and in-class readiness. However, over time, some students shifted towards passive use, relying on copy-paste behaviours. Engagement was shaped by learner autonomy, digital literacy, home support, and instructional scaffolding. Teacher reflections highlighted the need for guided prompts, closer monitoring of pre-class work, and more support for less independent learners. While ChatGPT offered cognitive benefits for vocabulary development, its effectiveness depended on how it was embedded within instructional design. The study emphasises the importance of teacher agency and intentional scaffolding to ensure that AI tools enhance, rather than hinder, vocabulary learning. Practical implications are discussed for educators, curriculum developers, and researchers seeking to implement AI in language education.