Enhancing Learning and Evaluation: AI as a Catalyst for Sustainable Education

Objective: This paper aims to explore how AI can support the achievement of SDG 4 by improving the accessibility and quality of education. The study’s objective is to design and implement an automated evaluation tool that ensures accuracy, efficiency, and fairness in assessments. Furthermore, the study also investigates the accuracy and reliability of an AI-based evaluation tool and compares it to conventional evaluation methods. The study designs and implements an AI-based assessment tool, IntelliScore, using pre-trained large language models — Llama3-70B and DeepSeek-R1-Distill-Llama-70B. IntelliScore is trained based on evaluation parameters. Theoretical approach: the theoretical backdrop of this study is built upon Bloom’s taxonomy theory. Method: for empirical purposes, the variation between the results of IntelliScore and conventional grading methods has been analyzed. Result: the results reveal that IntelliScore can be integrated into education to achieve various pedagogical benefits, including assessing students’ work and predicting their academic performance. Conclusion: by leveraging AI to streamline assessments and support teachers and students, the overall education system can be enhanced, contributing significantly to the achievement of SDG 4’s goals of equitable, inclusive, and quality education for all. The findings from this paper have implications for teachers, students, and policymakers. Overall, this study provides a valuable resource for researchers and practitioners in the field of education to deploy AI tools in their lifelong learning process.