ChemMindLab: An Intelligent Web-Based Platform for the Integrated Development of Computational and AI Thinking

Abstract AI for Science (AI4S) education has become an important direction in chemistry education as artificial intelligence increasingly shapes the formulation, screening, and validation of molecular and materials problems. ChemMindLab was developed as a web-based instructional platform that uses AI4S-oriented molecular candidate-screening activities to cultivate interdisciplinary learners with both computational and AI thinking. Through two representative instructional activities, the platform introduces students to research practices resembling authentic AI4S materials discovery: comparison of quantum mechanical (QM) calculations with machine-learning-potential (MLP) predictions for a molecular structure, followed by extension to isomeric candidate-set screening. A context-aware large language model (LLM) assistant supports explanation, operational guidance, and troubleshooting. Preliminary classroom use with undergraduate and graduate chemistry students suggests that ChemMindLab can support guided activities based on the two instructional workflows.