The Korea Institute of Curriculum and Evaluation (KICE) is considering introducing subjective and essay-type assessments to the College Scholastic Ability Test (CSAT), and the possibility of introducing AI automatic grading technology for large-scale answer processing has also been raised. However, researchers at the Korea Institute of Curriculum and Evaluation pointed out that current AI technology is insufficient to guarantee fairness and reliability in grading. They argued that the process of calculating scores with AI is like a 'black box' and opaque, and there is a possibility that it will rely on superficial characteristics such as the number of characters rather than the logic or reasoning of the answer. Therefore, they recommended using AI not as the main grader but as an auxiliary tool to support teachers' judgment and strengthening human review.
