AI-Driven Feedback and Learner Psychology: Evaluating Gemini and TalkAI in Enhancing Argumentative Writing Among IELTS Candidates

This study investigates the impact of artificial intelligence–mediated feedback on IELTS candidates’ argumentative writing performance and related psychological variables. Using a mixed-methods quasi-experimental design, ninety adult EFL learners were assigned to two AI-assisted groups (Gemini and TalkAI) or a traditional teacher-feedback control group over a six-week instructional period. Quantitative data were collected through IELTS Writing Task 2 pre- and post-tests, the Academic Motivation Scale (AMS), and the Self-Efficacy for Writing Scale (SEWS), complemented by semi-structured learner interviews. Statistical analyses revealed significant improvements in writing performance, academic motivation, and writing self-efficacy among participants receiving AI-based feedback, while changes in the control group were minimal. Qualitative findings further indicated that learners perceived AI feedback as timely, supportive, and conducive to increased autonomy and confidence. Overall, the study provides empirical evidence on the role of AI-driven feedback in enhancing both cognitive and affective dimensions of writing development in high-stakes EFL test preparation contexts.