Machine Learning-Enabled Automated Assessment and Grading in Education: A Systematic Literature Review of Techniques, Accuracy, and Fairness

Abstract

This research aims to synthesize recent literature on machine learning-enabled automated assessment and grading in education, with emphasis on techniques, accuracy, fairness, and responsible implementation. A systematic literature review was conducted on 27 English-language peer-reviewed journal articles published between 2019 and 2025. The review focused on the major applications and evaluation issues of automated grading, including essay scoring, short-answer assessment, programming assessment, feedback generation, validity, explainability, and fairness. The findings show that natural language processing, transformer-based models, supervised machine learning, hybrid feature-based approaches, code execution, and testing frameworks are commonly used. Automated grading is more suitable for structured or semi-structured tasks, while complex writing, creativity, reasoning, and programming style require human moderation. Institutions should adopt automated grading as a supervised decision-support tool supported by validation, fairness auditing, transparency, appeal procedures, and teacher oversight. The paper contributes a focused synthesis linking techniques, accuracy, fairness, and governance in automated educational assessment.