Artificial intelligence in higher education and graduate employability: cognitive engagement, self-efficacy, and career adaptability in the AI era

Artificial intelligence (AI), particularly generative AI, is increasingly used in higher education, but its contribution to graduate employability cannot be inferred from access to AI or frequency of use alone. This critical narrative review examines how AI-supported learning may relate to employability through processes discussed in educational psychology and career development. Searches of Web of Science, Scopus, ERIC, and PubMed covered database inception to 1 July 2026, with supplementary searches completed on 2 October 2026. The literature was synthesized interpretively and critically, with attention to the type of outcome measured and the extent to which findings support the proposed relationships. The synthesis focuses on cognitive engagement and self-regulated learning, self-efficacy and motivational regulation, and career adaptability. Three tensions organize the discussion: capability augmentation and cognitive offloading, career agency and AI-related insecurity, and resource access and capability differences. Included studies report associations or short-term changes in outcomes such as engagement, AI literacy, self-efficacy, career adaptability, and perceived employability. Direct evidence linking AI-supported learning to later employment outcomes remains limited. The proposed framework should therefore be read as a set of theory-informed pathways for future testing, rather than as an established causal or temporal sequence. Implications for assessment, disciplinary AI literacy, career education, institutional support, and future research are discussed.