Associations between deep learning vs. surface learning styles with both AI competence and personal enhancement by AI

Christian Montag, Jon D. Elhai,
Associations between deep learning vs. surface learning styles with both AI competence and personal enhancement by AI,
Learning and Motivation,
Volume 92,
2025,
102182,
ISSN 0023-9690,
https://doi.org/10.1016/j.lmot.2025.102182.
(https://www.sciencedirect.com/science/article/pii/S002396902500089X)
Abstract: Societal discussion has recently focused on what degree people use AI for personal enhancement such as using their own cognitive functions more effectively. Discussion has also centered on whether AI use has detrimental consequences for a user’s cognitive abilities. This new debate needs to be supplemented by empirical research, whereas the present research aims to shed light on who has a positive view about using AI for personal enhancement in the context of individual differences in learning styles. For this reason, n = 266 German-participants from the general population were recruited. In addition, n = 283 US-participants with a student background were recruited. All participants filled in a range of questionnaires including a measure on deep/surface learning style, items related to self-determination-theory in the context of AI and the Personal Enhancement by AI Scale (PEAIS). In the German sample, we observed that greater self-reported competence in AI was associated with more personal enhancement by AI tendencies. This association was mediated by deep learning style. In the US student sample, we replicated this association between AI-competence and greater personal enhancement by AI, but deep learning style was unrelated to both AI competence and personal enhancement by AI in this sample. Interestingly, post-hoc analyses revealed that associations between the three variables (AI-competence, deep learning style, personal enhancement by AI) were visible with small to moderate effect sizes in the US male subsample; so the lack of association between deep learning style and personal enhancement by AI within the US female sample was a failed replication. The present study findings should be seen as preliminary, and we hope to encourage other researchers to conduct follow-up research in this important, new research area.
Keywords: Deep learning; Surface learning; Artificial intelligence; Self-determination theory