Speakers
Description
When technology is integrated into educational environments, chatbots, as intelligent technological tools for learning assistance, garner widespread attention. However, despite the potential of chatbots in educational settings, there is relatively limited research on their application in guiding writing activities. Although current iterations of generative AI (GenAI) chatbots, such as ChatGPT, provide feedback on writing, how it facilitates EFL writing is still under-explored. This project aims to explore the efficacy of integrating ChatGPT in EFL writing and examine how EFL learners benefit from the GenAI feedback.
Thirty EFL college learners from a fundamental English composition class took part in the study. The learners were engaged in two in-class essay writing, producing two initial essay drafts (one cause-effect and one comparison). The learners then (1) revised the drafts using GenAI (ChatGPT) generated feedback produced based on the instruction given by the researcher, and (2) wrote a learning log, reflecting on how they revised the drafts based on the AI-generated feedback. The initial and revised drafts were compared regarding the development of topic sentences, the relevance of supporting ideas, the usage of academic language, and idea transition. The learning logs were coded for reflection episodes on topic sentences, supporting ideas, language usage, and idea transition.
The initial analysis of the learning logs indicated that the learners were able to identify the majority of their writing problems through the GenAI (ChatGPT) -generated feedback, particularly the language problems. The revised drafts were also better in language quality compared to the initial ones. The results thus suggest the potential of integrating GenAI chatbots to guide and facilitate English writing in an EFL learning environment.
| Keywords | customerised generative AI chatbots, EFL writing skills, learning autonomy, language complexity |
|---|
