Assessing the Capabilities and Limitations of Generative Artificial Intelligence in Solving College Chemistry Problems:Research Report Based on Student Group Practices
LI He-Xuan1, GUO Jin-Xin2**
1. School of Basic Medical Sciences, Shandong University, Jinan 250012, China; 2. College of Chemistry and Chemical Engineering, Shandong University, Jinan 250100, China
Abstract The rapid advancement of generative artificial intelligence(GAI) presents both unprecedented opportunities and challenges for education.To evaluate the potential of GAI in college chemistry education,this study selected 29 open-ended chemistry questions spanning different levels of Bloom’s cognitive taxonomy to assess the problem-solving capabilities of five major GAI models:ChatGPT-4.0,Kimi Assistant,Doubao,ERNIE Bot,and iFlytek Spark.The results reveal notable disparities in the models’ capabilities,especially in areas such as chemical mechanism elucidation,chemical calculation and synthesis analysis.These findings provide insights and recommendations for GAI application and future optimization in chemistry education.The study’s core innovations include two breakthroughs:First,it systematically compared the adaptability of five GAI models in teaching scenarios,addressing a gap in research that primarily focused on single models.Second,grounded in project-based learning(PBL) among first-year medical students,it establishes an integrated “AI assessment—critical thinking cultivation” framework supported by empirical evidence from the learner’s perspective.This student-centered approach not only enhanced students’ deeper understanding of chemistry but also reduced the risks associated with overreliance on AI,offering methodological support and practical guidance for the deeper integration of artificial intelligence into chemistry education.
LI He-Xuan, GUO Jin-Xin. Assessing the Capabilities and Limitations of Generative Artificial Intelligence in Solving College Chemistry Problems:Research Report Based on Student Group Practices[J]. Chinese Journal of Chemical Education, 2026, 47(8): 116-122.
