<![CDATA[This study examines the transformation of English language learning through Artificial Intelligence (AI)-assisted Task-Based Learning (TBL) and its effects on university students’ learning achievement and learner autonomy. A quasi-experimental pretest–posttest control-group design was prepared for 72 EFL university students divided into an experimental class (n = 36) and a control class (n = 36). The experimental class completed six authentic language-task cycles supported by generative AI for brainstorming, interaction rehearsal, language feedback, and reflective revision, whereas the control class implemented conventional Task-Based Learning without generative AI assistance. English learning achievement was measured through an integrated achievement test and task-performance rubric, while learner autonomy was assessed using a 24-item scale covering goal setting, strategic learning, self-monitoring, and self-evaluation. The draft statistical model indicates stronger post-intervention achievement and autonomy in the AI-assisted TBL group, with medium-to-large adjusted effects. The pedagogical mechanism is interpreted as a combination of authentic task orientation, immediate adaptive support, repeated opportunities for revision, and structured reflection. However, AI was most beneficial when its role was restricted to scaffolding rather than replacing students’ decision making. The study proposes a human-in-the-loop AI-TBL sequence consisting of pre-task orientation, AI-supported task performance, verification, peer negotiation, production, and reflective evaluation. These findings highlight that the transformation of English instruction depends not merely on access to AI tools but on the pedagogical architecture that preserves learner agency, critical verification, and accountable language production.]]>
Copyrights © 2026
