Implementation of AI in career counselling for university students

A systematic review

Dados Bibliográficos

ID22163207
AutoresYerbol Sarmurzin (0000-0003-3142-0545, Karaganda Buketov University), Askar Azhenov (0000-0003-0295-883X, Kazakh National Medical University), Ainash Kudysheva (0000-0001-8897-7497, University of Arts), Bayan Ualikhanova (0000-0001-7679-380X, Kajaani University of Applied Sciences), Indira Saparbekova (Kajaani University of Applied Sciences), Akan Besbayev (0009-0008-2810-1766, M.Auezov South Kazakhstan State University)
Ano2026
Volume11
Data de publicação2026-03-13
Open AccessSim
TipoARTICLE
PeriódicoFrontiers in Education (JOURNAL)
Identificadores do periódicoISSN: 2504-284X • E-ISSN: 2504-284X
EditoraFrontiers Media SA (PUBLISHER • CH)
DOI10.3389/feduc.2026.1787689
OpenAlexW7135170745
IdiomaEN
Referências citadas35

Artificial intelligence (AI) is transforming university career services, shifting from traditional predictive algorithms to interactive generative agents. This transition raises critical questions about student agency and the digital platformisation of career guidance. In this systematic review, we examined empirical evaluations and technical architectures of AI implementations in higher education career counselling from 2015 to 2025. Following PRISMA 2020 guidelines, we identified 43 studies across Web of Science, Scopus, and ERIC. Methodologically, we appraised study quality using the Mixed Methods Appraisal Tool (MMAT) and conducted a thematic narrative synthesis. We identified three main AI implementation typologies: (1) Generative Chatbots offering scalable 24/7 support, though limited by user trust barriers and a lack of perceived empathy; (2) Predictive Analytics forecasting employability, which require transparent, explainable models; and (3) Recommender Ecosystems that match students to personalized career pathways using institutional data. Interpreting these empirical findings through Career Construction Theory and Platformisation, we highlight a tension between technological efficiency and student agency. While AI expands information access, over-reliance on opaque algorithms introduces the risk of “automation bias,” where students passively accept recommendations rather than actively constructing their vocational identities. We conclude that AI should augment, rather than replace, human counsellors within a hybrid support model. Future research should prioritize longitudinal evaluations of AI interventions, algorithmic audits of recommender platforms, and the development of Explainable AI (XAI) to foster critical AI literacy

Analytics · Audit · Cognitive Information Processing · Empirical research · Implementation · Narrative · Recommender system · Vocational education · AI in Service Interactions · Artificial Intelligence in Healthcare and Education · Digital Mental Health Interventions

Ferramentas

Abrir DOIOpen Access