AI literacy alone is not enough: Student AI readiness and career adaptability in business and management education

Synthesis: Testa, Apuzzo, and Pittaway (2026) investigate how AI-related competencies contribute to career adaptability in business and management education. Surveying 339 university students in economics, management, and business programs in Italy, they employ a moderated mediation model examining the relationships among AI literacy, AI readiness, AI Self-Efficacy, and career adapt-abilities. Findings show AI readiness mediates the relationship between AI literacy and career adapt-abilities — with the indirect effect varying across levels of AI self-efficacy — and that AI self-efficacy positively moderates the literacy–readiness link while also directly associating with career adapt-abilities. The study positions AI readiness as a central mechanism linking AI-related learning to career-relevant outcomes, arguing that AI literacy alone does not directly translate into career adaptability.

Key Findings

• AI readiness mediates the relationship between AI literacy and career adapt-abilities — literacy alone does not directly translate into career outcomes; competencies become career-relevant when translated into readiness to engage with AI in applied settings.
• AI literacy predicts AI readiness (β = 0.148, p = .005), and AI self-efficacy also predicts AI readiness (β = 0.178, p = .005) — the model predicting AI readiness was significant (F(3,335) = 38.64, p < .001, R² = 0.257).
• AI self-efficacy moderates the literacy–readiness link: the interaction was significant (β = 0.299, p < .001); the effect of AI literacy on AI readiness was not significant at low self-efficacy (β = −0.067) but positive at medium (β = 0.123) and high (β = 0.327) levels.
• AI self-efficacy shows a direct positive association with career adapt-abilities (β = 0.255, p < .001) — a distinct motivational resource beyond literacy — and also interacts with AI literacy (β = 0.236, p < .001) in predicting career outcomes.
• Moderated mediation confirmed: the indirect effect of AI literacy on career adapt-abilities through AI readiness was not significant at low self-efficacy (β = −0.012) but positive and significant at medium (β = 0.022) and high (β = 0.057) levels, with a significant index of moderated mediation (index = 0.052).
• The full model predicting career adapt-abilities explained 40% of variance (F(4,334) = 55.75, p < .001, R² = 0.4004), with AI literacy (β = 0.120, p = .008) and AI readiness (β = 0.175, p < .001) as significant predictors.

Study Design & Method

This survey study collected data from 339 university students enrolled in economics, management, and business-related degree programs (including Computer and Data Science for Economics and Corporate Communication) at Italian higher education institutions, using an online questionnaire (Microsoft Forms) distributed between March and June 2025. Students had substantive exposure to AI within their curricula (group work, laboratories, or project-based assignments involving AI). Constructs were measured with validated scales: AI literacy, AI readiness (operationalized at the individual level as motivational and cognitive readiness), AI self-efficacy (Wang & Chuang, 2024), and career adaptability (Career Adapt-Abilities Scale – Short Form, measuring concern, control, curiosity, confidence). Items were translated from English with a back-translation procedure and pre-tested with 17 students. Hypotheses were tested using linear regression and moderated mediation analysis (PROCESS Model 8, version 4.0, Hayes 2022) in SPSS, with all variables standardized. Common-method bias was mitigated procedurally and statistically, and all scales demonstrated adequate reliability (Cronbach's α = 0.817–0.829; CR = 0.742–0.860; AVE = 0.602–0.769).

What this means for practice

• Learners. Do not treat literacy as the finish line: the effect of AI literacy on career adapt-abilities ran through AI readiness, so the competencies only became career-relevant once they translated into readiness to use AI in applied settings.
• Learners. Seek supervised, iterative experimentation with AI tools rather than one-off exposure — the indirect effect was not significant at low AI self-efficacy (β = −0.012) but positive at medium (β = 0.022) and high (β = 0.057) levels.
• Learners. Treat confidence as a distinct resource: AI self-efficacy had a direct association with career adapt-abilities (β = 0.255, p < .001) beyond the readiness pathway, so self-efficacy work is not a by-product of skills training.
• Instructors. Keep literacy, readiness, and self-efficacy separate in course design and assessment instead of reporting one aggregate AI-skills index, because the three played different roles in this model.
• Instructors. Assess applied judgment — problem framing, verification of outputs, responsible use — alongside technical understanding, and thread AI literacy across courses and progression points rather than confining it to a single module.

Limitations

• The design is cross-sectional and captures AI literacy, readiness, and career adaptability at an early, anticipatory stage of professional development, so it does not observe how these resources are enacted in stable organizational roles.
• The empirical setting is Italian higher education, where educational norms, labor-market expectations, and interpretations of AI may differ from other national contexts.
• All data were collected through a single survey instrument, though the authors report that common-method bias was mitigated.
• The sample is confined to economics/management/business students; the authors call for assessment across a broader spectrum of disciplines and cohorts and extension to early-career professionals.

Citation

Testa, M., Apuzzo, A., & Pittaway, L. (2026). AI literacy alone is not enough: Student AI readiness and career adaptability in business and management education. The International Journal of Management Education, 24, 101394.