The role of artificial intelligence in improving student mathematics instruction

Abstract

The article examines the complex nature of student’s learning difficulties in mathematics at a technical university. Based on a review of domestic and international literature, a multi-component structure of these barriers is identified, including cognitive, operational, and affective components. The author developed and tested a diagnostic questionnaire that allowed for a differentiated assessment of subject gaps, typical errors, and student’s subjective perceptions of mathematics. The empirical results revealed that the key difficulties are "fear of the blank page" (60 %), arithmetic errors (40 %), and misunderstanding of terminology (35 %), while student’s self-efficacy remains relatively high. The study substantiates the potential of artificial intelligence as a personalized support tool capable of adapting learning materials, providing feedback, and reducing emotional tension. Based on the findings, a typology of student AI interaction scenarios and directions for further methodological development are proposed.