Developing better questions: meta-problem tuning strategies based on conversational AI and MToM evaluation

Abstract

AI-based educational reforms have become mandatory. Education authorities in multiple nations have integrated AI-assisted learning competencies with a dialectical approach to enhance student competency. Teachers ought to analyze conversational AI’s advantages and disadvantages from several angles, evaluate AI capabilities sensibly, raise students’ problem-solving skills, and increase the potency of meta-problems. Based on MToM, we evaluate several conversational AI products from China, the US, and Russia in this study and find that they still have shortcomings in terms of behavioral cognition, semantic recognition, and other areas in Chinese. This demonstrates that conversational AI is still merely a tool at this point, lacking social interaction and human cognitive qualities. Student users should align their perceptions of AI product capabilities and develop their discernment skills, particularly those based on problem awareness, in order to improve meta-problem capacities. This approach allows teachers to gradually alter their teaching tactics as AI evolves