Ethical dilemmas of generative AI in English language teaching and governance pathways: review of practice barriers and optimization strategies

Abstract

Generative artificial intelligence (GenAI) is increasingly used in English language teaching, enabling personalized support, rapid development of learning materials, simulation of communicative situations, and faster feedback. At the same time, widespread classroom adoption intensifies ethical risks and governance challenges. This paper systematizes the major ethical dilemmas of GenAI use in English language education, including blurred boundaries of academic integrity and weakened trust in assessment, reduced learner autonomy and agency due to overreliance on ready-made outputs, threats to privacy and security of educational data, algorithmic bias and cultural misrepresentation in generated content, and the potential amplification of digital inequality arising from unequal access to tools and skills. To address these issues, the paper proposes a multi-level governance approach that reframes GenAI as a supportive instrument rather than a substitute for learning, combines transparent rules and procedures with AI and ethical digital literacy for teachers and students, and integrates technical safeguards such as data minimization, risk auditing, and bias mitigation. The proposed pathway aims to support responsible integration of GenAI while preserving humanistic educational goals and the validity of learning outcomes.