Research Article | Volume 115 Issue 12 (2025) | Published in 2025-12-16
Explainable AI Frameworks for Detecting Bias in Educational Assessment Systems
-
ABSTRACT
This research examines the use of Explainable Artificial Intelligence (XAI) in educational assessment systems, focusing on its potential to detect and mitigate bias in automated grading and predictive analytics. The study aims to evaluate how XAI frameworks, such as SHAP, LIME, and counterfactual explanations, enhance transparency, fairness, and accountability in student evaluation processes. The research problem arises from the growing reliance on AI in education, where opaque decision-making can lead to unintended discrimination, inaccuracies, or unequal treatment of students from diverse backgrounds. Key research questions include: How do XAI frameworks reveal sources of bias in educational AI systems? To what extent can these frameworks support equitable assessment practices? The study hypothesizes that implementing XAI in educational assessment improves fairness and interpretability, allowing educators to make more informed decisions while reducing the risk of bias and enhancing student trust in automated evaluation tools.
Keywords: Explainable Artificial Intelligence (XAI); AI Frameworks;
Bias Detection; Algorithmic Bias; Educational Assessment;
Assessment Systems. -
-
المراجع
[1] Ahmed, R., & Porter, D. (2021). Accountability and transparency in educational AI systems. Journal of Educational Technology, 15(2), 90–95.
[2] Barton, K., & Cheung, L. (2020). Rule-based explainable AI in automated assessment. International Journal of Artificial Intelligence in Education, 30(1), 50–60.
[3] Binns, R. (2018). Fairness in machine learning: Lessons from ethical governance. In Proceedings of the 2018 Conference on AI Ethics, 1–15.
[4] Bradley, S., & Hassan, M. (2022). Automated essay scoring and explainability: Ensuring fairness in writing assessment. Educational Assessment Review, 28(3), 100–110.
[5] Crawford, K., & Calo, R. (2016). There is a blind spot in AI research. Nature, 538(7625), 112–115.
[6] Daniels, P. (2021). Counterfactual explanations for bias detection in educational AI. AI in Education Journal, 23(2), 45–50.
[7] Dawson, T. (2022). Visualization tools for bias detection in educational AI. Computers & Education, 182, 115–125.
[8] Doshi-Velez, F., & Kim, B. (2017). Towards a rigorous science of interpretable machine learning. arXiv preprint arXiv:1702.08608, 1–13.
[9] Floridi, L. (2019). Artificial intelligence, education, and ethics: Transparency and accountability. Philosophy & Technology, 32(1), 40–50.
[10] Foster, J., & Li, H. (2022). SHAP explanations in educational assessment systems. Journal of AI Research in Education, 33(2), 100–110.
[11] Garrison, M., & Patel, S. (2021). Predictive analytics and fairness in educational institutions. Education Data Science, 14(1), 55–65.
[12] Hamada, Y. (2023). Explainable AI for language proficiency assessment. Journal of Language Learning Technologies, 11(2), 125–130.
[13] Harrison, K. (2022). Detecting bias in educational AI using explainable frameworks. International Journal of Educational Technology, 19(1), 58–62.
[14] Holstein, K., Wortman Vaughan, J., Daumé, H., Dudik, M., & Wallach, H. (2019). Improving fairness and interpretability in AI-driven educational platforms. Proceedings of the ACM Conference on Learning at Scale, 50–60.
[15] Jobin, A. (2019). The global landscape of AI ethics guidelines. Nature Machine Intelligence, 1(9), 54–59.
[16] Kleinberg, J., Mullainathan, S., & Raghavan, M. (2018). Inherent trade-offs in algorithmic fairness. Proceedings of the 8th Innovations in Theoretical Computer Science, 140–150.
[17] Kowalski, P., & Ahmed, F. (2022). Fairness-aware explainable AI frameworks for education. Educational AI Review, 16(3), 135–145.
[18] Lee, C. (2023). Explainable AI for short-answer and open-response grading. International Journal of AI in Education, 35(1), 90–95.
[19] Lopez, R., & Chang, Y. (2023). Student-centered assessment with explainable AI. Computers & Education, 187, 140–150.
-
Article history_ar
Received : Oct 22, 2025
Revised : Oct 24, 2025
Accepted : Nov 30, 2025
-
Authors Affiliations_ar
-
Ethics declarations_ar
Acknowledgment None Author Contribution All authors contributed equally to the main contributor to this paper. All authors read and approved the final paper. Conflicts of Interest “The authors declare no conflict of interest.” Funding “This research received no external funding”
