Impact of Meta AI-Generated Corrective Feedback in the Writing Classroom

Previous research has explored the impact of corrective feedback provided by human instructors on the writing skills development of ESL/EFL (English as a second/foreign language) learners. There also has been a growing trend towards employing corrective feedback generated by Meta AI for the same pedagogical purposes. However, no studies to date have examined the effect of such feedback on reducing grammatical errors in English writing among language learners., Therefore, the present study aims to examine the effects of three distinct types of corrective feedback, direct, indirect, and metalinguistic generated by Meta AI in conjunction with the WhatsApp mobile application, on grammatical errors in the written English of first-year undergraduate university students. The study employed random sampling to select four sections of undergraduates, comprising a total of 227 students. Three sections were assigned as the experimental group and one as the control group. Section A (N=59) received direct corrective feedback, Section B (N=53) received indirect corrective feedback, Section C (N=58) received metalinguistic corrective feedback, and Section D acted as the control group (N=57). Pictures were given to the students to compose a story in English, totalling 300 words, to collect data from them through a pretest and post-test. The data were analysed in terms of morphological, syntactic, and orthographic errors, employing the theoretical framework of Corder’s (1974) and Dulay’s (1982) taxonomies. The frequency of errors was recorded, and a repeated-measures ANOVA (analysis of variance) was used to analyse the data in SPSS (version 26). The statistical analysis revealed that the mean of errors decreased in the post-test writings of each section. Significantly, the metalinguistic corrective feedback produced by Meta AI proved effective in reducing errors compared to both direct and indirect feedback methods, in addition to the control group. This study suggests that integrating AI-generated metalinguistic feedback into English learning and teaching curricula could enhance error correction and learning outcomes in higher education.

Corrective Feedback · English language · Statistical analysis · EFL/ESL Teaching and Learning · Second Language Acquisition and Learning · Technology-Enhanced Education Studies