Theoretical statistics education with ChatGPT

This study explores the potential applications of ChatGPT in theoretical statistics education.To this end, the problem-solving capabilities of two ChatGPT-based models, GPT-4o and o3-mini, were evaluated, and the effectiveness of personalized learning was examined through a GPT-based tutorbot implemented in theoretical statistics courses at Korea National Open University.The findings suggest that ChatGPT is highly useful for solving undergraduate-level theoretical statistics problems and can serve as a valuable tool for promoting self-directed learning.Nonetheless, limitations such as the possibility of errors and excessive reliance on the system were also identified.Consequently, when integrating ChatGPT into theoretical statistics education, it is imperative to adopt pedagogical strategies that preserve creativity and foster critical thinking.