The proliferation of generative AI tools like ChatGPT is transforming higher education, yet evidence of their impact on students' career readiness remains limited. Existing studies often rely on localized samples and emphasize usage frequency over underlying mechanisms. This study investigates how ChatGPT usage relates to students' perceived employability, focusing on perceived proficiency (e.g. writing, communication) and analytical skills (e.g. critical thinking, problem-solving) as mediators.
Drawing on Self-Perception Theory and Human Capital Theory, with the Technology Acceptance Model providing a conceptual starting point for understanding technology engagement, we use structural equation modeling (SEM) to analyze survey data from 21,716 students across 100 countries, examining associations among ChatGPT use, perceived proficiency and analytical skill development, and perceived employability, with mediation through the two skills.
ChatGPT use is positively associated with both skill domains; however, only perceived proficiency is significantly associated with perceived employability. The indirect effect via proficiency is significant, whereas the indirect effect via analytical skills is not, and the direct path from ChatGPT use to employability is nonsignificant, consistent with a full-mediation model. Multi-group analyses indicate stronger effects for students from low- and lower-middle-income countries and for female students, with no substantive differences by field of study.
This study provides large-scale, theory-driven evidence linking ChatGPT use to perceived skill development and employability. These findings suggest that structured, critical engagement with AI tools may support skill development without displacing deeper learning. It offers critical insights for higher education policy and pedagogy in the GenAI era.
