生成式人工智能辅助初一学生解决数学问题的表现研究OA

Generative artificial intelligence(GAI)holds significant potential for application in mathematics education.To observe students'performance in GAI-assisted mathematics learning,this study selected eight seventh-grade students as participants.It analyzed their independent interactions with GAI while solving congruence problems based on three dimensions:the quantity,structure,and types of prompts used.Additionally,the problem-solving process was evaluated across four aspects:solution accuracy,reasoning methods,derivation processes,and symbolic notation.Data analysis revealed that during GAI-assisted learning,students demonstrated deficiencies in prompt design and superficial problem-solving approaches,with notable individual differences in human-AI interaction efficiency and effectiveness.After GAI-assisted learning,students exhibited poor knowledge transfer capability,degradation in mathematical expression skills,and superficial cognitive understanding.Accordingly,recommendations include:teachers should strengthen guidance and training on students'prompt design,deepen cognitive interventions in human-AI collaboration,implement tailored GAI-assisted learning activities,and adopt a cyclical"AI deconstruction-student reconstruction"learning model.