Evaluating Generative AI Tools’ Responses to Elimination and Nucleophilic Substitution Reactions Using Chemistry’s Triplet Model

The rapid development of generative artificial intelligence (AI) tools is increasingly evident in chemistry education. Understanding how these tools can be integrated into instruction, along with recognizing their advantages and limitations, is essential for designing effective learning environments. Additionally, adequately incorporating the interconnected macroscopic, submicroscopic, and symbolic levels—fundamental to the nature of chemistry—is critical for fostering deep and scientifically grounded learning. In this context, this study aims to investigate how elimination and nucleophilic substitution reactions, central topics in Organic Chemistry, are interpreted by different generative AI tools within the framework of chemistry’s triplet. As part of this case study, 15 questions related to these reaction types were posed to the generative AI tools “Gemini,” “Copilot,” “MagicSchool,” and “ChatGPT,” and each was additionally asked to explain a written SN1 reaction. Descriptive analysis revealed that, although all AI tools provided relatively adequate responses at the macroscopic level, their performance at the submicroscopic and symbolic levels was notably limited. The absence of curved arrow notation in illustrating reaction mechanism steps and the lack of three-dimensional structural representations required for the stereochemistry of reactions were identified as significant shortcomings. Furthermore, only Gemini was able to generate energy–reaction coordinate diagrams and provide examples of cyclic structures for E2 reactions. Furthermore, it was determined that each generative AI tool could explain reactions, even within the same reaction type, to varying degrees. Overall, these findings suggest that generative AI tools require further refinement, particularly to strengthen explanations at the submicroscopic and symbolic levels, for practical use in chemistry education.