When Learning Platforms Learn About Us: How AI Can Personalize Education at Scale

Makassar, Indonesia, November 1st, 2024/asds/ Imagine an online learning system that can quietly observe how students read materials, participate in discussions, complete quizzes, and manage assignments—and then use that information to understand how each student learns best. This idea is no longer science fiction. A recent scientific study shows how artificial intelligence (AI) can automatically detect students’ learning styles using data from online learning platforms, opening new opportunities for more adaptive, inclusive, and effective education.

The study, published in the International Journal of Advanced Computer Science and Applications, introduces an automated method to identify learning styles without relying on traditional questionnaires. For years, educators have used surveys based on models such as the Felder–Silverman Learning Style Model (FSLSM) to understand whether students prefer active or reflective learning, visual or verbal materials, sequential or global explanations. While useful, survey-based approaches are often time-consuming, subjective, and difficult to apply in large online classes.

The new approach takes a different path. Instead of asking students questions, the system learns from their behavior. By analyzing data already stored in a Learning Management System (LMS)—such as how long students spend on content, how often they revise quizzes, or how actively they participate in discussion forums—the researchers trained a machine learning model to recognize learning patterns automatically. The method combines a Support Vector Machine (SVM) with Recursive Feature Elimination (RFE), allowing the system to focus only on the most relevant indicators of learning behavior.

The results are striking. The AI model achieved high accuracy across all four FSLSM dimensions, with performance exceeding 95% in some cases. This means the system can reliably identify learning preferences at scale, helping educators move beyond one-size-fits-all teaching. With this insight, online courses can be designed to adapt content formats, learning paths, and feedback styles to match how students actually learn, not just how teachers assume they learn.

Beyond technical achievement, the implications of this work are deeply connected to the Sustainable Development Goals (SDGs). First, it supports SDG 4 (Quality Education) by promoting personalized and inclusive learning environments, especially in large-scale online education systems. When teaching adapts to learners, engagement improves and learning outcomes become more equitable. Second, it contributes to SDG 1 (No Poverty) by strengthening human capital through better education quality—an essential foundation for economic opportunity and long-term social mobility. Finally, the research reflects SDG 17 (Partnerships for the Goals), as it integrates knowledge from computer science, statistics, education, and data analytics to solve complex educational challenges.

In an era where digital learning is no longer optional but essential, this study shows that online platforms can do more than deliver content—they can understand learners. By responsibly using learning data and AI, education systems can become smarter, fairer, and more responsive. When technology learns how we learn, the future of education becomes not only digital, but truly human-centered.

http://dx.doi.org/10.14569/IJACSA.2024.0150661

Contact: Prof. Sri Astuti Thamrin

+6285298934579

tuti@unhas.ac.id