Study on the Improvement of Learners’ Academic Writing Competence Based on Multi-Round Automatic Feedback Supported by Generative Artificial Intelligence

This study explores the empowerment mechanisms and optimization paths of Generative Artificial Intelligence (AIGC) for academic writing competence. By constructing a human-machine collaboration framework covering the entire writing cycle, it proposes a four-dimensional interactive model of “Delay-Dialogue-Practice-Evaluation”: in the pre-writing phase, a three-stage cognitive intervention mechanism is adopted to avoid technology dependence; in the text generation phase, a hierarchical creation model and a Socratic dialogue system are developed; in the evaluation phase, a multi-dimensional diagnosis and metacognitive reflection mechanism is established. This study innovatively integrates process writing theory with the characteristics of AI technology, providing a sustainable development plan for academic writing education in the intelligent era that balances efficiency improvement and ethical constraints.