Career-related psychological cognitions and career exploration behavior among Chinese college students: statistical indirect pathways through generative artificial intelligence use

Generative artificial intelligence is changing how college students access career information, understand career opportunities, and prepare for job searching. Drawing on Social Cognitive Career Theory and the Technology Acceptance Model, this study used cross-sectional survey data to examine the relationships among career-related psychological cognitions, generative AI use, and career exploration behavior, and to test statistical indirect relationships through generative AI use. The sample consisted of 1,445 upper-year undergraduates from five universities in eastern China. Confirmatory factor analysis, structural equation modeling, and bootstrap analyses were conducted. In the primary model, career self-efficacy and perceived usefulness of generative AI were significantly and positively associated with both generative AI use and career exploration behavior, and generative AI use was also significantly and positively associated with career exploration behavior. The path from career self-efficacy to generative AI use was statistically significant but small ( β = 0.082), as was the corresponding statistical indirect relationship with career exploration behavior; neither was significant in the Sample B-only sensitivity analysis. Perceived usefulness of generative AI showed a significant statistical indirect relationship with career exploration behavior through generative AI use in the primary model. Career insecurity showed a weak but significant positive bivariate correlation with generative AI use ( r = 0.095, p < 0.001); however, after career self-efficacy and perceived usefulness of generative AI were included simultaneously in the structural model, the path from career insecurity to generative AI use was close to zero and nonsignificant ( β = 0.000, p = 0.989). Career insecurity showed neither a significant structural path to career exploration behavior nor a significant statistical indirect relationship with it through generative AI use. Overall, the findings indicate that career self-efficacy and perceived usefulness of generative AI were associated with career exploration behavior in the primary model, whereas career insecurity did not show a significant independent structural path after the other predictors were considered simultaneously. Because the data were cross-sectional and self-reported, the indirect relationships should be interpreted as statistical associations rather than evidence of causal relationships.