Document Type
Publication - Article
Comparing AI and Human Self-assessments in Memorization Performance
Department
Computing
Date of Activity
10-29-2025
Abstract
While there is some research that explores the effects of using artificial intelligence in a classroom setting, there has been very little exploration into how machine-learning-reliant methods are compared to traditional study methods, specifically in the area of memorizing complex topics. This study analyzes the effects of replacing a student’s perceived learning efficacy of a subject with a natural language understanding (NLU) model. We compared the effects of studying with a spaced-repetition application that utilizes an NLU model to calculate a user’s understanding of material against the effects of studying with a spaced-repetition model that does not use NLU. We used our results to determine if replacing the self-assessment component of flashcard studying applications with an NLU model leads to better memorization and retention. We found that while both approaches produced similar results, the non-NLU approach produced slightly higher and more consistent memorization outcomes than the NLU approach.
Recommended Citation
Ermer, M., Moses-Pakkianathan, A., Alférez, G.H. (2026). Comparing AI and Human Self-assessments in Memorization Performance. In: Arai, K. (eds) Proceedings of the Future Technologies Conference (FTC) 2025, Volume 4. FTC 2025. Lecture Notes in Networks and Systems, vol 1678. Springer, Cham. https://doi.org/10.1007/978-3-032-07992-3_6