skills gap

Background: Fresh graduates face an increasingly difficult school-to-work transition, while artificial intelligence (AI) has become a routine study and job-search companion; yet how the intention to use AI translates into employability, and through which mechanisms, remains unclear. Objective: Grounded in Social Cognitive Theory, this study tests whether behavioral intention to use AI is associated with self-perceived graduate employability directly and indirectly through career adaptability and job search self-efficacy, and whether AI self-efficacy and AI trust moderate these paths. Methods: A cross-sectional online survey of 612 Indonesian fresh graduates who use AI for career preparation was analyzed with partial least squares structural equation modeling (PLS-SEM). Results: AI use intention positively predicted career adaptability (β = 0.305) and job search self-efficacy (β = 0.249), and both resources predicted self-perceived employability. The indirect paths through both resources were significant, and a direct effect of AI use intention on employability remained, indicating partial mediation. AI self-efficacy and AI trust showed significant main effects on their resources, while the interaction-related indirect effects were not significant. Conclusion: AI use intention is associated with self-perceived employability through career adaptability and job search self-efficacy. The cross-sectional findings do not establish causality. Educational institutions and career-service units may use these findings to inform structured AI-supported occupational exploration, skills-gap analysis, résumé refinement, and interview rehearsal, while encouraging critical verification of AI outputs.

Students in vocational practical courses start with uneven skills, so an instructional model has to cope with that gap without losing sight of individual accountability. This quasi-experimental pretest–posttest study compared Team Assisted Individualization (TAI) with conventional instruction in a Fabric Embellishment Techniques course of a Diploma III Fashion Design program. Two intact parallel classes of 13 students each (N = 26) took part in the second semester of academic year 2024–2025, and a lottery decided which class received TAI. Practical performance was scored on a 0–100 scale before and after instruction. The TAI class rose from 72.38 to 88.85 and the conventional class from 70.77 to 81.31, mean gains of 16.46 and 10.54 points. The classes did not differ significantly at pretest (p = .231), although the TAI class started 1.62 points higher. At posttest the difference was 7.54 points, t(24) = 4.42, p < .001, and the difference in gains was 5.92 points, t(24) = 6.48, p < .001. Because each condition was delivered to a single intact class, these student-level tests cannot separate the teaching model from class-specific influences, and the results are best treated as preliminary evidence.