Abstract The rapid development of artificial intelligence generated content (AIGC) presents new opportunities and challenges for education. To assess the potential application of generative AI in high school chemistry education, this study evaluated the problem-solving abilities of two major generative AI models, ChatGPT and iFlytek Spark, on a set of chemistry questions. Their performances were compared with those of students from key high schools who selected chemistry as one of subjects for college entrance examination. The results indicated that the accuracy of the both AI models was only higher than that of a small group of students, with students generally outperforming both models. Further analysis of error types revealed that generative AI encountered issues such as misunderstandings of concepts and insufficient information extraction during problem-solving. Based on these findings, the study provides insights and recommendations for the use of generative AI in chemistry education and its future optimization and development.
CHEN Ling-Fang, ZHAN Xiao-Jun, XUE Song, DU Lin-Cun. Exploring Effectiveness on the Ability of Generative Artificial Intelligence to Solve High School Chemistry Problems[J]. Chinese Journal of Chemical Education, 2025, 46(9): 69-74.
