Development and Implementation of an Automatic Grading Quiz Application Using OpenAI Chat API

Abstract

　In this study, I developed an automatic grading system using OpenAI’s Chat API as a means to promptly grade the answers obtained from short-answer format questions posed during lectures to check students' understanding, and introduced it in a university programming class. The results showed that it is possible to grade with a certain level of accuracy by having AI evaluate the answers based on their similarity to the model answers. However, it was also revealed that there were some mistakes in marking. In the future, I aim to improve the efficiency and accuracy of the system by improving the prompts and combining it with other answer analysis methods such as keyword matching.