Understanding EFL students ' use of self-made AI chatbots as personalized writing assistance tools: A mixed methods study

This study aimed to explore English as a foreign language (EFL) students' use of self-made retrieval augmented generation (RAG) chatbots to enhance their learning to write. In the study, 69 Chinese undergraduate students participated in a workshop focused on creating chatbots, using Poe, that can assist with their writing processes. Multiple data sources were collected, including chatbots built by students, essays students wrote using their chatbots, students' responses to pre- and post-workshop questionnaires, and written reflections. The findings revealed that students developed chatbots for various purposes, such as assisting with idea generation, producing writing outlines, and identifying grammatical and spelling errors. Students made various requests, including assistance, customization, and translation, during their interactions with chatbots. Moreover, the use of self-made chatbots had a positive impact on students' writing motivation. It resulted in clearer writing goals, increased writing confidence, reinforced writing beliefs, and a more positive attitude towards writing. This study contributes to a deeper understanding of chatbots as pedagogical tools that enhance personalized language learning for students. By leveraging self-made chatbots, students can receive tailored support for their specific writing needs, leading to improved motivation.