Related Experiment Videos
Learning-centred use of generative AI and later academic functioning: a baseline-adjusted three-wave panel study.
Yang Zhao1, Jian Chen2, Wei Dai3
1School of Law, Southwest University of Science and Technology, Mianyang, Sichuan, China.
Frontiers in Psychology
|July 16, 2026
Summary
Learning-centred use of generative artificial intelligence (GenAI) enhances students' academic functioning by fostering self-regulated learning and academic self-efficacy. This approach, focused on task completion and verification, proves more valuable than simply measuring GenAI adoption rates.
Area of Science:
- Educational Psychology
- Artificial Intelligence in Education
- Higher Education Studies
Background:
- Generative artificial intelligence (GenAI) is increasingly prevalent in higher education.
- Existing metrics like adoption rates or general attitudes do not fully capture how students leverage GenAI for academic benefit.
- A need exists to understand specific GenAI usage patterns that support learning and academic success.
Purpose of the Study:
- To define and investigate "learning-centred use" (LCU) of GenAI in academic work.
- To examine the relationship between LCU and subsequent academic functioning (procrastination, engagement).
- To test the mediating roles of self-regulated learning (SRL) and academic self-efficacy (ASE) in the LCU-academic functioning pathway.
Main Methods:
- Analysis of a three-wave panel dataset from 1,200 university students in China.
- Measurement of LCU at baseline, followed by SRL and ASE at T2, and academic procrastination and learning engagement at T3.
- Application of baseline-adjusted composite-score path models, controlling for covariates and using inverse probability weighting for robustness checks.
Main Results:
- Learning-centred use (LCU) of GenAI positively predicted both self-regulated learning (SRL) and academic self-efficacy (ASE).
- SRL and ASE, in turn, predicted reduced academic procrastination and increased learning engagement.
- Indirect effects of LCU on academic outcomes were fully mediated by SRL and ASE, with no significant direct paths remaining.
Conclusions:
- The academic value of GenAI is realized through its integration into learning-centred regulatory practices, not merely through frequency of use.
- LCU acts as a crucial behavioral condition linking GenAI use to improved academic functioning via enhanced SRL and ASE.
- Higher education should guide students towards task-focused, verification-oriented, and non-substitutive GenAI use to support academic development.