Answers are grounded in this paper and cite their sources.
Background Writing anxiety remains a major obstacle to second language (L2) writing, particularly among vocational college students with limited English proficiency. Although artificial intelligence (AI)-powered chatbots have shown considerable potential in supporting language learning, their role in alleviating writing anxiety has received insufficient empirical attention. Methods This study examined whether integrating an AI-based chatbot into English writing instruction was associated with changes in English writing anxiety among Chinese vocational college students. A quasi-experimental design was employed with 120 students from two intact classes, with one class designated as the experimental group and the other as the control group. Over eight weeks, the experimental group received blended writing instruction incorporating Doubao, an AI-based chatbot, whereas the control group received conventional writing instruction. Writing anxiety was measured before and after the intervention across five dimensions: test anxiety, self-doubt, somatic anxiety, time anxiety, and avoidance behavior. Semi-structured interviews were also conducted with ten students from the experimental group to explore their learning experiences and perceptions of chatbot-assisted writing. Results The quantitative analyses showed that the experimental group experienced greater reductions in somatic anxiety and avoidance behavior than the control group, whereas no statistically significant between-group differences were found in test anxiety, self-doubt, or time anxiety. Qualitative evidence indicated that participants perceived the chatbot as a supportive and non-judgmental learning companion and reported reduced emotional pressure and greater writing confidence. Participants also described actively seeking help from the chatbot, checking its suggestions, and revising their writing during the writing process. Conclusion These findings suggest that, under the instructional conditions examined, AI chatbot-supported blended writing instruction was associated with alleviations in some dimensions of writing anxiety. The findings should be interpreted in the context of the broader instructional condition, which combined chatbot support with regular classroom instruction and writing activities. As writing performance was not directly assessed, the study focused on changes in writing anxiety rather than writing proficiency or writing quality. Further research with larger samples, more closely matched instructional conditions, and longer follow-up periods is needed to further examine these relationships.
No takes yet. Share an insight, caveat, or question.
Bao et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: