Answers are grounded in this paper and cite their sources.
This study explored the role of ChatGPT4o mini as an AI teacher assessor (ATA) in Chinese English-as-a-foreign-language (EFL) writing classrooms. It first used univariate and multivariate generalizability (G-) theory to compare the holistic and analytic scoring provided by ATA with that of experienced human teacher assessors (HTAs) across 25 EFL essays. The results indicated that ATA demonstrates reliability comparable to HTAs, with high reliability coefficients observed in both holistic and analytic scoring. It then compared the quality of feedback provided by ATA with that of HTAs and found that ATA provided more consistent and comprehensive feedback across the domains of language, content and organization. It finally interviewed HTAs and found that they held a cautiously optimistic attitude towards the adoption of ATA, recognizing its reliability and efficiency, but also expressing concerns about accessibility, ethical implications and the depth of its feedback compared to HTA feedback. Implications for Chinese EFL HTAs are discussed.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: