Creating new materials with the help of artificial intelligence has become routine. Models can generate millions of variants in minutes, but in practice, this rarely leads to truly useful innovations for fields like computer chip manufacturing or rocket engines. The problem is that many of these generated materials turn out to be unstable and unsuitable for real-world use.
A new approach developed by researchers solves this problem at its root. Artificial intelligence now not only generates but also predicts how stable a proposed material will be. The algorithm analyzes structure and properties, identifying potential issues even before costly synthesis experiments begin. This means engineers receive not just a list of ideas, but guaranteed viable options.
This approach directly impacts development costs. It is estimated that up to 90% of the effort in creating new materials was spent on sifting through unsuccessful, unstable samples. Now, this percentage can be minimized.
This opens the door for faster integration of advanced materials into high-tech products, from more efficient cooling systems for electronics to components for extreme operating conditions. Scientists have effectively taught AI not just to dream, but to soberly assess its dreams, which in itself is quite significant.