Application of Artificial Intelligence Models in Standardized Tests for Optimizing Academic Performance in Higher Education

Abstract

Introduction: Although AI has shown potential to predict academic outcomes, design personalized learning programs, and support academic guidance, there are significant challenges such as the need for high-quality data, interpretability issues of some models, and the risk of perpetuating existing biases. The object of this systematic review explores the use of artificial intelligence in the educational field, specifically in the context of standardized tests. Methodology: To achieve this, a comprehensive review of the scientific literature is conducted following the PRISMA statement guidelines, with a sample of 17 articles published between 2019 and 2023 in journals indexed in Scopus. Results: The most used predictive models in these studies were found to be: Artificial Neural Networks, Decision Trees, Support Vector Machines, and Random Forest, Discussions: identifying benefits that include optimizing academic performance, personalizing learning, and improving educational decision-making. Conclusions: It is concluded that AI has great potential to improve the measurement of educational quality, but it is crucial to address these limitations and ethical considerations to ensure its effective and responsible application in the educational field.