AI-Supported Feedback and Cognitive Load in EFL Writing: Evidence from Learners' Development

Generative artificial intelligence has created new opportunities for supporting second language (L۲) writing. Although tools such as Chat GPT are increasingly used in EFL instruction, their cognitive effects during writing remain underexplored. Drawing on Cognitive Load Theory and metacognitive monitoring, this study examined whether AI-assisted feedback could improve writing accuracy and complexity while reducing perceived mental effort. Forty EFL learners aged ۱۶-۲۰ participated in a ten-week classroom intervention using ChatGPT for guided feedback on writing tasks. A single-group pre-test/post-test design was employed. Writing samples were analyzed for accuracy and lexical and syntactic complexity, and learners completed a cognitive load questionnaire. Results from paired-samples t-tests showed significant gains in accuracy, moderate improvement in complexity, and lower perceived mental effort after the intervention. These findings suggest that AI-assisted feedback can act as a cognitive scaffold that supports revision and more efficient attention allocation during L۲ writing.

کلیدواژه ها:

Artificial intelligence (AI) ، cognitive load ، writing development ، metacognitive monitoring ، foreign language