TEACHERS’ DECISION-MAKING IN ADAPTING AI-GENERATED ACADEMIC TEXTS FOR TASK-BASED LANGUAGE TEACHING (TBLT) IN UZBEKISTAN

Abstract

This thesis analyzes teachers’ pedagogical decision-making in selecting and adapting artificial intelligence-generated academic texts for Task-Based Language Teaching (TBLT) in the context of teaching English as a foreign language (EFL) in Uzbekistan. The study considers texts generated by generative AI tools not as ready-to-use teaching materials, but as resources that require linguistic, content-related, and methodological evaluation and adaptation by teachers to suit specific learning tasks. Particular attention is paid to teachers’ decisions based on lexical and grammatical complexity, content accuracy and reliability, students’ language proficiency, the characteristics of academic discourse, the communicative purpose of the task, and relevance to the local educational context. The study also systematizes adaptation strategies such as simplifying and shortening texts, replacing lexical items, contextualizing content, and transforming texts into TBLT tasks. The analysis demonstrates that artificial intelligence does not automatically guarantee pedagogically appropriate materials and that the quality of the final learning materials depends on teachers’ professional judgment, linguistic competence, and methodological decisions.