The Learning Assurance Frontier: A Capacity-Constrained Theory of Generative AI, Knowledge Verification, and Independent Learning Transfer
Keywords:
generative artificial intelligence, independent learning transfer, epistemic verification, educational assessment, cognitive offloading, learning assurance frontierAbstract
Generative artificial intelligence (GenAI) can improve the speed and polish of assessed work without necessarily strengthening knowledge that learners can independently reproduce. Existing accounts of cognitive offloading and responsible AI identify this risk, but provide limited formal guidance for allocating verification effort when AI output volume rises faster than learner and educator capacity. This article develops the Learning Assurance Frontier (LAF), a capacity-constrained theory of AI-assisted learning that distinguishes production, verification, and delayed independent transfer. Using analytic modelling, constrained optimisation, and transparent numerical scenarios, it derives conditional propositions: (i) the fraction of AI-supported work that can be fully checked declines once verification demand exceeds finite review capacity; (ii) an explicitly specified assurance-adjusted educational objective has a unique, capacity-dependent optimal AI-assistance intensity; and (iii) unequal review resources can create assurance disparities even when students have equal access to AI tools. Additional results establish consequence-weighted allocation of limited review time and a distributionally robust safety margin for uncertain verification demand, both under explicit assumptions. In illustrative, non-empirical scenarios, increasing weekly verification capacity from 1.8 to 3.0 hours shifts the model-optimal assistance intensity from approximately 0.35 to 0.71. The article translates these conditional results into an assessment architecture incorporating evidence checks, self-explanation, delayed unaided transfer, and equity safeguards. It specifies an experimentally testable research programme, while explicitly avoiding any claim that the illustrative parameters are measured educational effects.
References
1. Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122(26), e2422633122. https://doi.org/10.1073/pnas.2422633122
2. Buçinca, Z., Malaya, M. B., & Gajos, K. Z. (2021). To trust or to think: Cognitive forcing functions can reduce overreliance on AI in AI-assisted decision-making. Proceedings of the ACM on Human-Computer Interaction, 5(CSCW1), Article 188. https://doi.org/10.1145/3449287
3. Chen, S., & Cheung, A. C. K. (2025). Effect of generative artificial intelligence on university students learning outcomes: A systematic review and meta-analysis. Educational Research Review, 49, Article 100737. https://doi.org/10.1016/j.edurev.2025.100737
4. Chi, M. T. H., & Wylie, R. (2014). The ICAP framework: Linking cognitive engagement to active learning outcomes. Educational Psychologist, 49(4), 219–243.
https://doi.org/10.1080/00461520.2014.965823
5. Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students’ learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14(1), 4–58.
https://doi.org/10.1177/1529100612453266
6. Kapur, M. (2014). Productive failure in learning math. Cognitive Science, 38(5), 1008–1022.
https://doi.org/10.1111/cogs.12107
7. Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier, E., Krusche, S., Kutyniok, G., Michaeli, T., Nerdel, C., Pfeffer, J., Poquet, O., Sailer, M., Schmidt, A., Seidel, T., ... Kasneci, G. (2023). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual Differences, 103, 102274. https://doi.org/10.1016/j.lindif.2023.102274
8. Long, D., & Magerko, B. (2020). What is AI literacy? Competencies and design considerations. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems (pp. 1–16). Association for Computing Machinery. https://doi.org/10.1145/3313831.3376727
9. Miao, F., & Holmes, W. (2023). Guidance for generative AI in education and research. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000386693
10. Noy, S., & Zhang, W. (2023). Experimental evidence on the productivity effects of generative artificial intelligence. Science, 381(6654), 187–192. https://doi.org/10.1126/science.adh2586
11. OECD. (2026). OECD Digital Education Outlook 2026: Exploring effective uses of generative AI in education. OECD Publishing. https://doi.org/10.1787/062a7394-en
12. Risko, E. F., & Gilbert, S. J. (2016). Cognitive offloading. Trends in Cognitive Sciences, 20(9), 676–688. https://doi.org/10.1016/j.tics.2016.07.002
13. Roediger, H. L., III, & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249–255. https://doi.org/10.1111/j.1467-9280 .2006.01693.x
14. Soderstrom, N. C., & Bjork, R. A. (2015). Learning versus performance: An integrative review. Perspectives on Psychological Science, 10(2), 176–199. https://doi.org/10.1177/1745691615569000
15. Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285. https://doi.org/10.1207/s15516709cog1202_4
16. Tan, K. H. (2026a). The cognitive sovereignty threshold: An interdisciplinary humanities and sciences framework for human agency in generative AI-assisted knowledge work. International Journal of Research Publication and Reviews, 7(7), 2445–2454. https://doi.org/10.55248/gengpi.07.0726.2117
17. Tan, K. H. (2026b). Improving critical thinking skills among adult learners through structured reflective dialogue: An action research study in a university social sciences course in Singapore. International Journal of Research Publication and Reviews, 7(6), 5322–5329.
https://doi.org/10.55248/gengpi.07.0626.17a03
18. Tan, K. H. (2026c). The synthetic consensus trap: Correlated AI errors, verification overload, and the mathematics of institutional epistemic cascades. ResearchGate.
https://doi.org/10.5281/zenodo.22009669
19. Tan, K. H. (2026d). The verification bottleneck in AI-accelerated knowledge work: Queueing epistemics, verification capacity, and adaptive human oversight. Open Access Journal of Multidisciplinary Research, 2(9), 1–17. https://doi.org/10.47760/OAJMR.2026.v02i09.001
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Author(s)
This work is licensed under a Creative Commons Attribution 4.0 International License.
Articles published in the Journal of Advanced Multidisciplinary Studies (JAMS) are licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0), unless otherwise stated. Authors retain copyright of their work and grant JAMS the right of first publication.
