A silent drag: Does artificial intelligence lower adolescents' academic performance?

As a disruptive technology, artificial intelligence (AI) has permeated nearly every sphere of social life, yet its role in empowering basic education remains underexplored. Drawing on cognitive load theory and expected human capital investment theory, this study examines how AI development affects adolescents' academic performance using matched data from educational statistics, smart home data, and the China Family Panel Studies. The results show that AI development significantly reduces adolescents' academic performance in both educational and daily-life contexts. This finding is robust to the use of the number of local government education officials in the late Qing dynasty as an instrumental variable, as well as a series of additional robustness checks. Mediation analysis indicates that individual cognitive ability and family educational expectations are key channels through which AI impairs adolescents' academic performance. Heterogeneity analysis further suggests that this negative effect is more pronounced among urban youth, lower-education groups, and regions with lower human capital premiums and denser off-campus tutoring industries. In addition, strong family educational concern can mitigate the adverse impact of AI on adolescents' academic performance. This study provides empirical evidence for understanding the mechanisms through which artificial intelligence influences primary education, and offers implications for the standardized governance of AI applications in primary education and for the formulation of family education support policies.