Now the age has opened in which artificial intelligence (AI) imagines and predicts the structures of new materials like a person. Going beyond a tool that merely calculates, it works together as a researcher's "second brain," from generating ideas to experimental validation.
A research team led by Professor Hong Seung-beom in the Department of Materials Science and Engineering at the Korea Advanced Institute of Science and Technology (KAIST) said on the 26th that, together with Drexel University, Northwestern University, the University of Chicago, and the University of Tennessee, it comprehensively analyzed the evolution of AI. The results were published in the international journal ACS Nano in Jul.
The team divided the materials research process into three stages—discovery, development, and optimization—and specified the role AI plays at each stage. AI first recommends promising candidates from among countless materials, reduces trial and error in the experimental process, and adjusts conditions on its own to find the optimal outcome.
The team analyzed that the latest technologies—such as Generative AI, graph neural networks, and transformer models—are turning AI from a simple calculation tool into a thinking researcher. AI learns the principles of physics and chemistry on its own to imagine new materials and carries out the entire process together from proposing ideas to experimental validation.
It also introduced cases of "autonomous laboratories," in which AI itself draws up experimental plans and robots carry out the experiments, and an "AI-based catalyst search platform." In this system, AI designs the experimental conditions, and robots automatically synthesize catalysts and analyze the results. This technology, which can speed up research, can also be applied to other fields such as batteries and energy materials.
However, the team noted that AI's predictions are not always the correct answer. It said tasks remain to be solved, including imbalances in data quality, difficulties in interpreting results, and the integration of disparate data. Therefore, it emphasized that going forward, technologies must advance in tandem so that AI understands physical principles and researchers can transparently verify the process.
Professor Hong Seung-beom said, "This study shows that AI is establishing itself not just as a tool but as a new language and way of thinking in materials science and engineering," adding, "The roadmap presented here will offer important direction to researchers in fields such as batteries, semiconductors, and energy materials."
References
ACS Nano (2025), DOI: https://doi.org/10.1021/acsnano.5c04200