Who Should Grade My Work? Student Perspectives on Transparent AI-Assisted Writing Assessment in Higher Education

Recent research has confirmed the feasibility of using AI in large-scale writing evaluation. A study based on real exam data proposed and verified a “Human-in-the-Loop” scoring framework. This strategy involves AI for initial evaluation and human experts for review of key cases, effectively reducing manual workload while maintaining moderate to high levels of scoring consistency. Another empirical study of students at Saudi universities found that although students accepted and used feedback generated by AI for text revision, they generally viewed AI as a helpful auxiliary tool, while considering human teachers as decision-makers with ultimate “evaluation authority.” Both studies indicate that when introducing AI-assisted scoring, it is necessary to combine carefully designed human-machine supervision processes to balance efficiency and accuracy, and clearly define the role boundaries between feedback utility and evaluation authority.