Advancing materials discovery through artificial intelligence

Abstract

Artificial intelligence (AI) is transforming materials science by accelerating the design, synthesis, and characterization of novel materials. This review highlights how AI, including machine learning, deep learning, and generative models, is reshaping the discovery pipeline. AI-driven approaches enable rapid property prediction, inverse design, and simulation of complex systems such as nanomaterials and solid-state materials, often matching the accuracy of ab initio methods at a fraction of the computational cost. Machine-learning-based force fields provide efficient and transferable models for large-scale simulations, while explainable AI improves transparency and physical interpretability. In synthesis, AI supports synthesis planning, reaction optimization, and the development of autonomous laboratories capable of real-time feedback and adaptive experimentation. Tools originally developed for organic molecules are increasingly adapted for complex materials, including those with structural, thermodynamic, or kinetic constraints. AI also advances in situ characterization, automating tasks such as spectral interpretation and defect identification. Despite rapid progress, challenges remain in model generalizability, standardized data formats, experimental validation, and energy efficiency. The review underscores the importance of hybrid approaches combining physical knowledge with data-driven models and calls for open-access datasets also including negative experiments and ethical frameworks to ensure responsible deployment. Future directions include modular AI systems, improved human-AI collaboration, integration with techno-economic analysis, and field-deployable robotics. By aligning computational innovation with practical implementation, AI is poised to drive scalable, sustainable, and interpretable materials discovery, turning autonomous experimentation into a powerful engine for scientific advancement.

Topics & Concepts