From AI Use to Pedagogical Legitimacy: Teacher Agency and Cross-Cultural Negotiation in GenAI-Mediated English Language Teaching in Thailand

Generative artificial intelligence (GenAI) is increasingly used in English language teaching (ELT), yet its educational value depends on how teachers evaluate and regulate its use within particular classroom and institutional contexts. This interpretive, multisite qualitative comparative interview study examined how English language teachers in Thailand made GenAI-supported practices educationally acceptable and professionally defensible. Semi-structured interviews were conducted with 100 teachers: 50 Thai and 50 international teachers representing diverse professional, linguistic, and educational backgrounds. Reflexive thematic analysis examined shared concerns, differing emphases, within-group variation, and divergent cases. A preliminary Techno-Lingual Synergy Framework informed the study as a set of sensitizing concepts rather than a predetermined coding structure. Five interconnected themes were developed: evaluative labor, provisional output, guided autonomy, negotiated classroom fit, and institutional governance. Participants valued GenAI for materials development, feedback, translation, differentiated instruction, and extended language practice, but these benefits depended on teachers’ capacity to verify output, adjust materials to learner needs, preserve evidence of learning, establish assessment boundaries, and restrict inappropriate use. Differences between Thai and international teachers reflected overlapping professional emphases rather than fixed cultural orientations and were often better explained by learner characteristics, institutional conditions, professional responsibilities, and experience teaching in Thailand. The findings refine the framework by conceptualizing teacher agency as pedagogical governance and pedagogical legitimacy as the provisional process through which AI-supported practices become educationally acceptable for particular purposes and contexts. Within the settings represented, responsible GenAI integration depended on the interaction of situated teacher judgment and institutional governance.