Experimental evaluation of Large Language Models for in-class learning experience customization

Abstract

This paper explores the utilization of Large Language Models (LLMs) for personalized education in Secondary Education, focusing on motivation and personalization. It uses the GPT 3.5 model by OpenAI to generate tailored exercises and examines their impact on student motivation and academic performance. The study highlights the positive correlation between motivation and performance, emphasizing the need to consider classroom dynamics and teacher-student relationships. While LLMs enhance educational content, they should complement, not replace, the teacher’s role. The research calls for further investigation into personalization’s impact on education, considering study duration and student samples for a more robust understanding.