Explainable Artificial Intelligence Based Academic Evaluation Systems: A Comprehensive Survey of Human in the Loop Decision Support and Multi Metric Writing Assessment

Artificial Intelligence (AI) has increasingly transformed academic evaluation by enabling automated assessment, intelligent feedback generation, plagiarism detection, and personalized learning support. The growing adoption of AI-driven evaluation systems has introduced new opportunities for improving the efficiency, scalability, and consistency of educational assessment, while also raising concerns regarding transparency, fairness, privacy, bias, and accountability. Recent advances in Explainable Artificial Intelligence (XAI) and large language models (LLMs) have further expanded the capabilities of AI-based academic evaluation by supporting more interpretable, context-aware, and human centered assessment processes. This survey presents a comprehensive review of AI based academic evaluation systems and organizes existing research across major methodological paradigms, including machine learning, deep learning, natural language processing, explainable AI, and LLM-based approaches. Beyond methodological categorization, we further examine how these approaches support diverse academic evaluation tasks, including automated grading, feedback generation, plagiarism detection, learning assessment, and personalized evaluation. Furthermore, we summarize commonly used datasets, evaluation metrics, explainability techniques, and Human-in-the-Loop decision-support strategies employed in existing studies. We also discuss key ethical and practical challenges associated with the deployment of AI in academic evaluation, with particular emphasis on fairness, bias, privacy, transparency, and accountability. Finally, we identify important open issues and future research directions. Overall, this survey provides a systematic reference for understanding the methodological development, practical applications, and responsible future of AI based academic evaluation.