<p>With the rapid development of artificial intelligence (AI), a new educational model where AI and teachers work together to improve learning outcomes has emerged. Although previous research has focused on comparing the effects of AI and teacher feedback, the division of labor between them has received little attention. Focusing on enhancing learners’ English argumentative writing, this paper examines the collaboration of AI and teacher in feedback provision and its impact on learners’ writing. Twenty-eight tenth-grade students participated in the study, using feedback from both their teachers and the AI tool, DeepL Write, to revise essays. The study compared the effects of AI-only and teacher + AI hybrid feedback using Coh-Metrix to analyze the writing quality across three drafts: the original (draft 1), the AI-assisted revision (draft 2), and the teacher-assisted revision on draft 2 (draft 3). The results revealed that the AI-only feedback was effective in lexical and grammatical aspects but had weaknesses in improving content selection, coherence and cohesion, which could be mitigated through the integrated use of teacher and AI feedback. Based on the findings, we propose a partially automated hybrid approach that allows AI writing tool to execute feedback provision at lower levels, while teachers focus on offering insights at higher levels of writing. The findings of this research will contribute to a better understanding of the mechanism of teacher and AI cooperation in feedback provision. Future research needs to explore students’ learning autonomy and assess the long-term impact of AI and teacher feedback integration in writing development.</p>