JOURNAL ARTICLE
Abstract: Educational Large Language Models (LLMs) outputs may contain factual errors, claims unsupported by course materials, citation defects, or cross-turn contradictions. These properties are related but not interchangeable: a correct claim may be absent from course materials, and course materials may themselves be mistaken. This paper presents a selective descriptive synthesis and proposes a course-grounded reporting framework. It organises five review dimensions into claim-level judgments (factual status, course support, and evidence integrity), a dialogue-level judgment (consistency), and a contextual interpretation of defect-related pedagogical risk. Learner outcomes are separate evidence requiring direct measures; causal attribution requires designs linking exposure to tutor outputs with measured outcomes. An eight-step procedure operationalises the review, and a cross-classification table makes factual status and course support explicit. The illustrative risk rubric is unvalidated and does not cover every source of instructional harm. Retrieval, abstention, and human oversight are safeguards to evaluate, not guarantees.
Keywords: large language models, educational AI, hallucination, learning outcomes, retrieval-augmented generation, reliability
Article Info: Received: 14 Oct 2026, Received in revised form: 10 Nov 2026, Accepted: 15 Nov 2026, Available online: 20 Nov 2026
Z. Ji et al., “Survey of Hallucination in Natural Language Generation,” ACM Computing Surveys, vol. 55, no. 12, Art. no. 248, pp. 1–38, 2023, doi: 10.1145/3571730.
K. U. Danyaro, S. Abdullahi, A. S. Abdallah, and H. Chiroma, “Hallucinations in Large Language Models for Education: Challenges and Mitigation,” International Journal of Teaching, Learning and Education, vol. 4, no. 6, pp. 13–19, 2025, doi: 10.22161/ijtle.4.6.2.
S. Abdullahi, Umama, K. U. Danyaro, H. Chiroma, A. Zakari, M. Abdullahi, S. H. Muhammad, and N. A. W. A. Zawawi, “Large Language Models Applications in Oil and Gas: Case Studies and Insights,” in 2025 International Conference on Artificial Intelligence for Sustainable Innovation (AI-SI), pp. 1–6, 2025, doi: 10.1109/AISI66213.2025.11341041.
H. Bastani, O. Bastani, A. Sungu, H. Ge, O. Kabakcı, and R. Mariman, “Generative AI without guardrails can harm learning: Evidence from high school mathematics,” Proceedings of the National Academy of Sciences, vol. 122, no. 26, Art. no. e2422633122, 2025, doi: 10.1073/pnas.2422633122.
Z. A. Pardos and S. Bhandari, “ChatGPT-generated help produces learning gains equivalent to human tutor-authored help on mathematics skills,” PLOS ONE, vol. 19, no. 5, Art. no. e0304013, 2024, doi: 10.1371/journal.pone.0304013.
Q. Jia, J. Cui, R. Xi, C. Liu, P. Rashid, R. Li, and E. Gehringer, “On assessing the faithfulness of LLM-generated feedback on student assignments,” in Proceedings of the 17th International Conference on Educational Data Mining, pp. 491–499, Atlanta, GA, USA, 2024, doi: 10.5281/zenodo.12729868.
R. Nemeth, A. T ´ atrai, M. Szab ´ o, P. T. Zaletnyik, and ´ A. Tam ´ asi, ´ “Exploring the use of retrieval-augmented generation models in higher education: A pilot study on artificial intelligence-based tutoring,” Social Sciences & Humanities Open, vol. 12, Art. no. 101751, 2025, doi: 10.1016/j.ssaho.2025.101751.
Y. Zhang, Z. Wang, C. Zhang, Y. Gao, W. Ma, and X. Zhai, “MisconceptionBench: Evaluating LLMs on Science Misconceptions and Student Response Distributions,” in Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pp. 16525–16539, Budapest, Hungary, 2026, doi: 10.18653/2026.emnlp-main.712.
A. Yerukola, J. D. Hwang, M. Zheng, J. Godsey, H. Kim, V. Pyatkin, J. Hu, and M. Sap, “XYBench: Can LLMs Respond Pragmatically to Queries with Misconceptions?” in Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pp. 60203– 60234, Budapest, Hungary, 2026, doi: 10.18653/2026.emnlp-main.2646.
T. Yasir, W. Li, S. Gilson, S. Tithi, X. Tian, and T. Barnes, “Confirming correct, missing the rest: LLM tutoring agents struggle where feedback matters most,” in Proceedings of the 21st Workshop on Innovative Use of NLP for Building Educational Applications, pp. 819–840, San Diego, CA, USA, 2026, doi: 10.18653/v1/2026.bea-1.56.
A. Nura and S. Abdullahi, “A systematic review of multi-depot vehicle routing problems,” Systematic Literature Review and Meta-Analysis Journal, vol. 3, no. 2, pp. 51–60, 2022, doi: 10.54480/slrm.v3i2.37.
P. Lewis et al., “Retrieval-augmented generation for knowledge-intensive NLP tasks,” in Advances in Neural Information Processing Systems, vol. 33, pp. 9459–9474, 2020.
R. Xie, D. Gopinath, D. Qiu, D. Lin, H. Sun, S. Potdar, and B. Dhingra, “Over-Searching in Search-Augmented Large Language Models,” in Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 7714–7739, Rabat, Morocco, 2026, doi: 10.18653/v1/2026.eacllong.361.
A. F. Hayes and K. Krippendorff, “Answering the call for a standard reliability measure for coding data,” Communication Methods and Measures, vol. 1, no. 1, pp. 77–89, 2007, doi: 10.1080/19312450709336664.
UNESCO, Guidance for generative AI in education and research. Paris, France: UNESCO, 2023.