- Bo Liu, Quanxi Chen
- DOI: 10.5281/zenodo.21497453
- GAS Journal of Education and Literature (GASJEL)
With
the deep penetration of generative artificial intelligence (GenAI) in higher
engineering education, its impacts on students’ higher-order thinking skills
and underlying mechanisms remain controversial. Grounded in the ICAP cognitive
engagement framework, this study integrates cognitive unloading theory,
dual-process theory, and cognitive load theory to construct a moderated
mediation model of “usage pattern—cognitive engagement—ability
outcome”. It examines how different AI usage patterns differentially
affect engineering undergraduates’ critical thinking and creative
self-efficacy, as well as the boundary moderating role of task complexity.
Adopting a two-wave longitudinal questionnaire design, this study surveyed 428
engineering undergraduates from a Double First-Class university in western
China, and conducted empirical tests via structural equation modeling and the
Bootstrap method. Results show that the theoretical model fits the data well.
Instrumental use positively predicts shallow cognitive engagement, which in
turn negatively affects critical thinking, presenting an “erosion
effect”. Exploratory use positively predicts deep cognitive engagement,
which significantly enhances creative self-efficacy, presenting an
“enablement effect”. Task complexity plays a positive moderating role
in the first half of both mediation paths: higher task complexity amplifies
both the cognitive erosion risk of instrumental use and the cognitive
enablement effect of exploratory use. This study confirms the dual effects of
AI in engineering education and reveals that depth of cognitive engagement is
the core regulatory mechanism of technology application effectiveness,
providing theoretical and practical references for reconstructing engineering
talent cultivation models in the AI era.
Keywords: Generative Artificial Intelligence, Instrumental Use, Exploratory Use, Cognitive Engagement, Critical Thinking, Creative Self-Efficacy, Engineering Education.
