Development Of An Artificial Intelligence-Based Essay Exam Assessment System

Abstract

This study aims to develop an Artificial Intelligence (AI)-based digital platform for automatic essay assessment of students at MTsN Filial Kartasura. The background of this research is the problem of traditional manual essay grading, which is time-consuming, less efficient, and potentially subjective. The research method used is Research and Development (R&D) with a 4D development model consisting of Define, Design, Develop, and Disseminate stages. The system was developed using PHP, HTML, CSS, and MySQL programming languages, and integrated with AI to analyze and assess student essays based on teacher-defined criteria. The results show that the AESAI system is capable of automatically scoring essays based on keyword relevance and answer quality, while also providing instant feedback. The Blackbox testing results indicate that all system features function properly. The material expert validation resulted in a score of 92.24%, categorized as very feasible, while the media expert evaluation obtained 93.21%, also categorized as very feasible. Furthermore, user testing involving 32 teachers produced an average score of 89.72%, categorized as very good. These results indicate that the developed system is well accepted by users. Therefore, the AESAI platform is proven to be feasible as an innovative solution for essay assessment. The system not only improves teachers’ efficiency in grading but also provides more objective, consistent, and transparent evaluation in digital learning environments.