AI-Based System for Detecting Students' Mathematical Errors and Reasoning Patterns: A Personalized, Accessible, and Human-Centered Approach to Transforming Math Education

Mathematics is often perceived by students—especially in the Arab world—as an intimidating and overly complex subject reserved only for the exceptionally gifted. This widespread misconception does not stem from a lack of student ability, but rather from traditional instructional methods that prioritize memorization, rigid procedures, and perfect exam scores over genuine understanding and cognitive development. As a result, many students develop deep-seated anxiety toward mathematics, which negatively impacts their academic confidence and future opportunities. This research introduces a fully integrated AI-based system designed to analyze students’ step-by-step solutions in mathematics, detect errors with high precision, classify the underlying reasoning patterns, and provide personalized, interactive, and visually guided feedback. Unlike classical teaching approaches that treat all learners as identical, the proposed system adapts to each student’s cognitive level, learning pace, and conceptual gaps. The system aims not only to correct mistakes, but to reshape the student’s relationship with mathematics by making the learning process intuitive, exploratory, and emotionally supportive. The significance of this work is particularly evident in countries such as Syria, where the national baccalaureate exam determines a student’s entire academic and professional trajectory. A strong foundation in mathematics can dramatically influence a student’s future, yet many capable learners fail to reach their potential due to fear, lack of resources, or ineffective teaching methods. While global researchers have explored modern pedagogical approaches that emphasize understanding over memorization, most advanced AI-powered educational tools remain paid and inaccessible to students in financially constrained regions. This research addresses this gap by offering a free, accessible, and adaptive AI system that democratizes high-quality math education. Early observations suggest that AI-driven step analysis can significantly reduce repeated errors, enhance conceptual understanding, and empower students to approach mathematics with confidence rather than fear. The proposed system represents a step toward reimagining math education as a personalized, equitable, and intellectually enriching experience for learners worldwide.