Abstract
The review is a synthesis of the recent research on the implementation of artificial intelligence in managing academic projects in higher education institutions, specifically semantic approaches to track academic integrity, evaluation of the originality of the research topic, and supervisor selection. The previous research has often addressed plagiarism detection, novelty assessment, and the recommendation as separate features; therefore, the current paper is an attempt to unify the methodology, current trends, and the existing challenges related to these tasks. Based on the guidelines of the Preferred Reporting Items of Systematic Reviews and Meta-Analyses (PRISMA) 2020, we chose 49 peer-reviewed studies that were published between 2021 and 2025 and were found in the large scientific databases. The paper combines a systematic review and bibliometric analysis, including keyword co-occurrence, co-citation analysis, and network analysis of co-authorships through VOSviewer. Four major thematic clusters were identified: (1) semantic integrity and plagiarism detection, (2) measuring the novelty and originality of the research topics, (3) intelligent recommendation and resource allocation systems, and (4) human-centred management and decision-support. In these clusters, lexical baselines are increasingly complemented or extended by embedding and transformer-based similarity models, and the most recent works use large language models to make higher-order reasoning and decisions. The literature is, however, fragmented, with an assortment of heterogeneous assessment procedures, limited use of real institution data, and scarce longitudinal validation. The results provide a comprehensive perspective on the methodology and research gaps, thus making it possible to build integrated academic management platforms, that combine integrity control, originality evaluation, and supervisory allocation