Speaker
Description
Writing in English is a challenging task. Grading papers often requires significant effort from teachers. Moreover, daily writing practice with computers can save a lot of study costs. For teachers, grading demands meticulousness and time spent reading papers, with results often subjective. With the advancement of AI, particularly in natural language processing, this study proposes a model for assessing English writing skills to evaluate the quality of AI grading. The research compares AI-graded data with human-graded data based on real data from UEH (University of Economics Ho Chi Minh City) examination. The results show a high correlation between AI and human grading, indicating that AI grading can assist teachers in making prompt decisions on grading papers. This article also highlights the limitations of AI grading and suggests some ways to overcome them. The author recommends using AI for learning purposes and suggests that in exams, it should only support grading rather than completely replacing human assessment.
