Scientists at the University of New Hampshire have used artificial intelligence to dramatically accelerate the search for advanced magnetic materials that could reduce dependence on rare earth elements, which are costly and difficult to source. By creating a massive, searchable database of 67,573 magnetic compounds, the research team — led by doctoral student Suman Itani — identified 25 previously unrecognised materials that remain magnetic at high temperatures, a key property needed for real-world applications. These magnets are essential components in devices like electric vehicles, medical tools, and power generators, making this discovery potentially transformative for technology manufacturing.
Traditionally, discovering new permanent magnets has been extremely slow and expensive because experimenting with millions of possible elemental combinations in the lab takes significant time and resources. The new AI system reads existing scientific literature to extract experimental data, trains models to assess magnetic properties, and predicts how materials behave at different temperatures. This approach enables researchers to explore vast areas of “materials space” far more quickly than conventional methods — opening up cheaper and more sustainable alternatives to rare earth magnets that currently dominate many technologies.
The searchable Northeast Materials Database created through this work makes it easier for scientists around the world to explore magnetic candidates and build on the research. Since rare earth elements are not only expensive but also largely imported from a few countries, finding domestic, high-temperature magnets could strengthen supply chains and help reduce geopolitical vulnerabilities in key industries like clean energy and electric transportation.
Beyond magnet discovery, the researchers see broader potential for the AI methods developed in this project. They believe the same technology could be used in higher education, for example to convert older scientific imagery into searchable formats for modern digital libraries. This underscores how AI isn’t just accelerating materials science — it’s also creating tools that may reshape how scientific knowledge is organised and accessed.
