Female secondary school students consistently demonstrate lower AI learning attitudes, literacy, and career interest than males, but AI literacy serves as the crucial bridge connecting positive learning attitudes to career aspirations regardless of gender.
Objective: This study aimed to investigate the relationships between secondary school students' AI learning attitude, AI literacy, and AI career interest, while examining gender differences in these relationships. The researchers sought to understand how these factors interact to influence students' engagement with AI education and their willingness to pursue AI-related careers.
Methods: The researchers employed a survey research design using stratified sampling to collect data from 622 secondary school students across 11 schools in China that offered AI courses. Three validated instruments were used: the AI Learning Attitude Survey (AI-LAS), AI Literacy Survey (AI-LS), and AI Career Interest Survey (AI-CIS). All instruments used 5-point Likert scales and were adapted from previously validated measures. The AI-LS was developed specifically for this study following established scale development procedures, incorporating expert reviews and pilot testing. Data analysis was conducted using structural equation modeling (SEM) with AMOS 28.0, including bootstrapping analysis for mediation effects and multi-group analysis for gender differences.
Key Findings: The study revealed several significant patterns. Female students consistently scored lower than male students across all three measures (AI learning attitude, AI literacy, and AI career interest), though effect sizes were small. Strong positive correlations were found between all variables: AI learning attitude and AI literacy (r=0.72), AI learning attitude and AI career interest (r=0.60), and AI literacy and AI career interest (r=0.69). The structural equation modeling revealed that AI learning attitude positively predicted both AI literacy (β=0.83) and career interest (β=0.70), while AI literacy also significantly predicted career interest (β=0.75). Most importantly, AI literacy fully mediated the relationship between AI learning attitude and career interest, accounting for 78.60% of the mediation effect. The direct effect of learning attitude on career interest became non-significant when AI literacy was included as a mediator, indicating complete mediation.
Implications: These findings have significant implications for AI education policy and practice. The mediating role of AI literacy suggests that simply fostering positive attitudes toward AI learning is insufficient for developing career interest; students must develop actual AI competencies and skills. This highlights the need for hands-on, practical AI education that goes beyond theoretical knowledge to include programming, problem-solving, and real-world application experiences. The persistent gender gaps indicate an urgent need for targeted interventions to support female students, including mentorship programs, female role models in curricula, and gender-sensitive teaching strategies. The study also suggests that early exposure to comprehensive AI education could help level the playing field by focusing on developing actual literacy rather than relying on pre-existing attitudes or stereotypes.
Limitations: The study has several notable limitations. The sample was drawn exclusively from schools already offering AI courses, potentially creating selection bias and limiting generalizability to schools without AI programs. The research relied entirely on self-reported measures, which may not accurately reflect actual AI competencies versus perceived abilities. The study was conducted only in China, limiting cross-cultural applicability. Additionally, the cross-sectional design prevents causal inferences about the relationships between variables. The focus on gender as the primary demographic variable meant other potentially important factors like socioeconomic status, prior technology exposure, or cultural background were not thoroughly examined.
Future Directions: The researchers recommend several areas for future investigation. Studies should include students from schools without AI courses to better understand the broader impact of AI education accessibility. Mixed-methods approaches incorporating objective AI literacy assessments, interviews, and classroom observations would provide more comprehensive understanding of these relationships. Cross-cultural studies across different educational systems and cultural contexts would help establish the generalizability of findings. Longitudinal research could track how AI literacy and career interest develop over time and examine the long-term career outcomes of students with different levels of AI education exposure. Finally, intervention studies testing specific strategies for reducing gender gaps and enhancing AI literacy could provide practical guidance for educators and policymakers.
Title and Authors: "Modeling the relationships between secondary school students' AI learning attitude, AI literacy and AI career interest" by Di Zhang, Hongwu Yang, Yanshan He, and Weitong Guo.
Published On: September 15, 2025 (published online)
Published By: Education and Information Technologies (Springer)