Socratic wisdom in the age of AI: a comparative study of ChatGPT and human tutors in enhancing critical thinking skills

Abstract

Introduction: This study investigates the perceptual differences between ChatGPT and human tutors in fostering critical thinking among students, highlighting the effectiveness of Socratic tutoring methodologies in modern educational contexts. Methods: Conducted with a sample of 230 university students in Taiwan, the research employed a mixed-methods approach, combining quantitative surveys and qualitative interviews. Results and discussion: Results reveal that while a significant portion of students appreciates the benefits of ChatGPT—namely its non-judgmental nature and accessibility—human tutors are acknowledged for their capability to provide tailored feedback and emotional support. Through threshold analysis conducted via a Multilayer Perceptron (MLP) model, the study identified key features affecting student preferences, emphasizing the need for a balanced integration of AI and human tutoring methods. Findings underscore the importance of developing hybrid educational models that leverage both the strengths of human facilitators and the efficiencies of AI tools to enhance student learning and critical thinking skills.