Generative artificial intelligence dependence and academic performance: serial mediation through attentional control and physical fitness and moderation by grade level
Year: 2026
Type: article
Study design: Observational Study
Abstract: Objective This study examined the association between generative artificial intelligence dependence (GenAID) and academic performance (AP) among adolescents, whether attentional control (AC) and physical fitness (PF) mediated this association individually and in sequence, and whether grade level moderated these pathways. Methods Participants were 1,461 Chinese adolescents. PF was assessed using standardized fitness tests; GenAID and AC were measured by questionnaires; and standardized scores in Chinese language, mathematics, and English served as indicators of AP. Structural equation modeling and related analyses were conducted using AMOS 24.0 and SPSS 26.0. Results GenAID was negatively correlated with AC, PF, and AP ( p < 0.01), whereas all other pairwise correlations were positive ( p < 0.01). All direct predictive paths among the variables in the serial mediation model and the predictive paths from the interaction terms to AP were statistically significant ( p < 0.01). GenAID significantly predicted lower AP through three indirect pathways: AC (95% CI [−0.067, −0.031]), PF (95% CI [−0.190, −0.003]), and AC followed by PF (95% CI [−0.009, −0.002]). Grade level significantly moderated the AC–AP and PF–AP paths: the standardized coefficients for Grade_AC_inter and Grade_PF_inter were −0.101 ( p < 0.001) and −0.164 ( p < 0.05), respectively. Conclusion GenAID directly predicted lower AP and also predicted lower AP indirectly through AC, PF, and AC followed by PF. The findings indicate interrelationships among GenAID, AC, PF, and AP. Grade level moderated the serial mediation model, with a weaker negative predictive relationship between GenAID and AP at higher grade levels.
Source: Frontiers in Psychology
Authors: Yu Su(Conceptualization, Formal analysis, Methodology, Software, Writing – original draft), Renren Zhang(Conceptualization, Supervision, Validation, Writing – review & editing), Pandeng Chen(Software, Visualization, Writing – original draft), Xiangyu Wang(Writing – review & editing), Yi He(Supervision, Validation, Writing – review & editing), Xingyu Zhang(Resources, Writing – review & editing)
Institution: Sichuan Normal University
Language: English
Subfield: Sociology and Political Science
Field: Social Sciences
Domain: Social Sciences
Open Access status: gold
Estimated APC paid: $3,801
OpenAlex ID: W7233926985
