Investigating How Learners' Attitudes Shape AI ‐Supported Argumentative Writing Processes and Outcomes

Background Artificial intelligence (AI), in particular, generative AI, plays an increasingly critical role in education. Studies have shown that learners' positive perception of generative AI‐assisted writing tools can benefit them in their writing processes and outcomes. However, despite the growing body of research on AI‐assisted writing, little is understood about how learner attitudes towards AI impact argumentative writing activities using tools like ChatGPT. Objectives This study investigates how learners' attitudes towards ChatGPT influence both quantitative outcomes (essay quality and perceived effectiveness) and qualitative aspects (students' knowledge construction processes and their perception of AI‐assisted argumentative writing) of writing activities. Methods This study employed a mixed‐method approach to examine how 51 undergraduate students (31 males, 20 females) at a university in South Korea completed argumentative essays using ChatGPT. We triangulated quantitative and qualitative data by (1) applying regression analyses and MANOVA to assess learner attitudes, (2) conducting a thematic analysis of participants' perception of key skills, benefits and challenges in AI‐assisted writing and (3) performing conversation analysis of learner‐ChatGPT interactions to evaluate knowledge construction practices. Results and Conclusions Findings revealed that positive attitudes towards ChatGPT significantly correlated with enhanced essay quality and perceived effectiveness. Moreover, students with high learner attitudes engaged in relatively more frequent and varied knowledge construction practices. However, some low‐attitude learners still produced high‐quality essays by employing strategic interactional practices such as negotiating epistemic stances and creating hypothetical scenarios, suggesting that healthy scepticism may foster critical engagement beneficial for argumentative writing. These findings provide important insights for understanding the role of learner attitudes in AI‐assisted writing and suggest extending TAM with ‘perceived dialogic affordances’—learners' perceptions of AI as a stance‐responsive interactional partner.