Yue Li, Shurui Wang, Jianping Wang, Liu Qian, Jin Zhang
This is a comprehensive review paper that systematically outlines the current state and future of AI fundamentally changing research methodologies in materials science.
There is a need to systematically analyze the role and impact of AI in materials science and to present future challenges and strategic directions for its advancement.
The literature is structured and analyzed around two main trends—task-specific AI and generalist AI—based on the materials discovery workflow. It examines how each type of AI is applied across various stages of materials research, such as hypothesis generation, experimental planning, and characterization.
The paper demonstrates that AI is innovating core stages of materials science, including hypothesis generation, experimental optimization, and autonomous laboratory operation. It clearly presents key challenges and development directions for the future ecosystem, such as knowledge representation, agent-based workflows, and human-AI collaboration, thereby providing a development roadmap for the field.