Speakers
Description
Many EFL and ESL teachers do not provide corrective feedback on L2 writing. Reasons include but are not limited to: 1. considerable time required; 2. unhelpful but necessary delay in returning corrective feedback; 3. concerns about comprehensibility of the feedback in the L2; and 4. concerns that students may simply disregard feedback. In recent years, automated corrective written feedback has become available in the form of commercial programs such as Grammarly and ProWritingAid. These programs potentially solve issues of long hours spent marking papers and the delay between students writing and receiving feedback. However, additional problems such as use of machine translation and copy-paste become an issue when moving from paper format to digital. And, crucially, programs such as the aforementioned usually offer feedback in the students’ L2, feedback that may not be comprehensible to all students.
With the advent of ChatGPT and other AI programs using large language models, it is now possible to provide automated written corrective feedback in the students’ L1. This presentation showcases results of an exploratory study using a ChatGPT-powered plugin currently in development for the popular Moodle LMS. The classroom intervention on which this study is based had four main steps: a) individual or collective brainstorming and online search for potentially useful vocabulary (3 minutes); b) subsequent free-writing activity in an online browser with copy and paste functions disabled (10 minutes); c) reading the ChatGPT feedback in L1 Japanese (2 minutes); d) completing paper grammar logs to process feedback (5 minutes).
Data has been compiled from students’ written submissions as well as the ChatGPT-generated feedback, students’ grammar logs, and survey instruments administered to students and teachers. The presenters aim to discuss concerns about accuracy and comprehensibility of the feedback in the L1, ease of use, what prompts work best, and student and teacher perspectives.
| Keywords | AI, ChatGPT, corrective feedback, writing |
|---|
