The review by Forero-Corba and Negre Bennasar maps, within a very recent time frame (2021–February 2023), how the AI/ML pairing is being used in education, based on 55 articles identified in Web of Science and Scopus following the PRISMA protocol.
The study deliberately narrows its corpus to English-language, open-access publications and organizes the landscape by educational level and geography. Research is distributed across 38 countries, with a dominant presence of the United States and a noteworthy shift toward primary and secondary education, where most of the reported applications are concentrated.
The central contribution is twofold. On the one hand, the authors inventory 33 techniques, with a clear predominance of supervised learning, and identify which algorithms most frequently appear in educational contexts (for example, Random Forest approaches, decision trees, and k-NN). On the other hand, they synthesize the main uses: prediction of academic performance and dropout, recommendation and academic guidance, analytics of teacher and student perceptions, robotics and virtual reality, AI literacy, and more specific cases such as cybersecurity, algorithmic fairness frameworks, and support for students with autism spectrum disorder.
The article stresses that the promise of these systems lies less in the “model” itself and more in data quality, labeling, and governance, as well as in the institutional capacity to translate predictions into educationally justifiable decisions.
In its conclusions, the review does not present AI/ML as a monolithic block. It distinguishes between classroom-level uses and management-level applications, notes the visible imprint of the post-pandemic context (with part of the corpus addressing COVID-19), and highlights a recurring bottleneck: teachers’ techno-pedagogical competence to interpret, constrain, and effectively leverage predictive tools.
The text also leaves open several tensions that call for further empirical work: bias introduced by publishing exclusively in English, reliance on only two databases, the scarcity of evidence related to diversity and special educational needs, and the need to ensure that “integrating AI” does not simply mean purchasing tools, but rather redesigning assessment, curriculum, and training according to ethical and traceable criteria.
---
How to Cite: Forero-Corba, W., & Negre Bennasar, F. (2024). Techniques and applications of Machine Learning and Artificial Intelligence in education: a systematic review. RIED-Revista Iberoamericana de Educación a Distancia, 27(1), 209–253. https://doi.org/10.5944/ried.27.1.37491
