Effectiveness of Artificial Intelligent Tutoring Systems for Learners with Limited Academic Proficiency: An Analytical Review in STEM Higher Education

Abstract

This analytical review explores the effectiveness of Artificial Intelligent Tutoring Systems (ITS) for academically underprepared learners in STEM higher education. The study synthesises recent advancements in generative AI and ITS, examining their potential to deliver personalised instruction, adaptive feedback, and scalable learning environments. It highlights the architectural components and operational mechanisms of ITS, evaluating their strengths in fostering academic improvement. The paper identifies key limitations including challenges in real-time adaptability, equitable access, and ethical data management, as well as concerns regarding the reliability of online assessments and the ability of ITS to replicate nuanced human guidance. The review calls for future research on emotion-aware computing, collaborative learning, explainable AI, emphasising the necessity for ethical, transparent, and accessible ITS solutions. Ultimately, the article argues that while ITS platforms hold significant promise for transforming STEM education and supporting underprepared learners, their success depend on continual innovation, robust evaluation, and a commitment to educational equity and excellence.