Practical challenges and optimization strategies for integrating generative artificial intelligence into English language teaching reform

Abstract

Generative artificial intelligence (GAI) offers new opportunities to reform English language teaching by accelerating lesson preparation, supporting personalized learning, and expanding feedback and assessment practices. However, classroom implementation reveals persistent practical challenges. This paper aims to systematize these challenges across four levels – technological, pedagogical, ethical-institutional, and learner/teacher development – and to propose an integrated optimization framework. The analysis shows that major obstacles include limited sensitivity to context and communicative intent, strong dependence of output quality on prompt formulation, difficulties in redefining the teacher’s role and embedding GAI into curricula, risks of academic misconduct and data privacy breaches, potential bias and unequal access, and the danger of reduced learner autonomy and critical thinking when GAI is used without clear boundaries. As a solution, the paper proposes a four-track optimization model – technology, pedagogy, ethics, and development – covering AI literacy and prompting skills, redesigned learning tasks and assessment with process evidence and protected components, transparent institutional rules for acceptable use and data protection, and targeted professional development for teachers alongside strategies to foster learner agency and reflective use of GAI.