TY  - JOUR
AU  - He, Chen
AU  - Yi, Wenlong
AU  - Yin, Hua
AU  - Zhao, Wei
PY  - 2026
DA  - 2026/07/24
TI  - From AI "Disabling" to "Enabling": An Empirical Study on Digital-Intelligent Learning among Students Majoring in Smart Agriculture
JO  - Journal of Digital Intelligence in Education
T2  - Journal of Digital Intelligence in Education
JF  - Journal of Digital Intelligence in Education
VL  - 1
IS  - 1
SP  - 17
EP  - 30
DO  - 10.62762/JDIE.2026.651948
UR  - https://www.icck.org/article/abs/JDIE.2026.651948
KW  - smart agriculture
KW  - digital-intelligent learning
KW  - large language models
KW  - structural equation modeling
AB  - Against the backdrop of deep AI-agriculture integration, smart agriculture has become a key direction for agricultural modernization, demanding higher digital-intelligent learning competence from university students. As a core competency for leveraging AI to engage in agricultural information technology learning and solve practical problems, the mechanisms underlying its formation remain underexplored. This study targets undergraduate and graduate students in smart agriculture-related majors, constructing a theoretical model based on the dual dimensions of external technological characteristics and internal learning processes. It systematically investigates how key factors—including the technological suitability of large language models (LLMs), users' information-processing capability, and their capacity to regulate LLM dependence—influence the formation of digital-intelligent learning competence. Based on 166 valid questionnaire responses, reliability and validity tests and factor analysis were conducted in SPSS, and Structural Equation Modeling (SEM) was performed in Amos. Results indicate that LLM technological suitability and learners' information-processing capability exert significant positive effects on digital-intelligent learning competence, while dependence-regulation capability serves as an important mediating variable. Further analysis reveals that LLMs do not directly enhance learning competence merely by improving efficiency; their enabling effect depends on learners' rational regulation of model usage and deep information processing. These findings uncover the critical pathways underlying the formation of digital-intelligent learning competence, providing theoretical foundations and practical guidance for universities to optimize talent cultivation models and enhance students' digital-intelligent learning competence.
SN  - pending
PB  - Institute of Central Computation and Knowledge
LA  - English
ER  - 

@article{He2026From,
  author = {Chen He and Wenlong Yi and Hua Yin and Wei Zhao},
  title = {From AI "Disabling" to "Enabling": An Empirical Study on Digital-Intelligent Learning among Students Majoring in Smart Agriculture},
  journal = {Journal of Digital Intelligence in Education},
  year = {2026},
  volume = {1},
  number = {1},
  pages = {17-30},
  doi = {10.62762/JDIE.2026.651948},
  url = {https://www.icck.org/article/abs/JDIE.2026.651948},
  abstract = {Against the backdrop of deep AI-agriculture integration, smart agriculture has become a key direction for agricultural modernization, demanding higher digital-intelligent learning competence from university students. As a core competency for leveraging AI to engage in agricultural information technology learning and solve practical problems, the mechanisms underlying its formation remain underexplored. This study targets undergraduate and graduate students in smart agriculture-related majors, constructing a theoretical model based on the dual dimensions of external technological characteristics and internal learning processes. It systematically investigates how key factors—including the technological suitability of large language models (LLMs), users' information-processing capability, and their capacity to regulate LLM dependence—influence the formation of digital-intelligent learning competence. Based on 166 valid questionnaire responses, reliability and validity tests and factor analysis were conducted in SPSS, and Structural Equation Modeling (SEM) was performed in Amos. Results indicate that LLM technological suitability and learners' information-processing capability exert significant positive effects on digital-intelligent learning competence, while dependence-regulation capability serves as an important mediating variable. Further analysis reveals that LLMs do not directly enhance learning competence merely by improving efficiency; their enabling effect depends on learners' rational regulation of model usage and deep information processing. These findings uncover the critical pathways underlying the formation of digital-intelligent learning competence, providing theoretical foundations and practical guidance for universities to optimize talent cultivation models and enhance students' digital-intelligent learning competence.},
  keywords = {smart agriculture, digital-intelligent learning, large language models, structural equation modeling},
  issn = {pending},
  publisher = {Institute of Central Computation and Knowledge}
}