What belongs here
New methods, meaningful results, useful software, and better data. We cover discovery, simulation, characterization, and autonomous experiments with an emphasis on what a materials researcher can learn or use.
Entirely AI-generated
Material Intelligence is an experiment in a purely AI-generated news outlet. AI selects the stories, summarizes the sources, and writes the articles. The reporting and commentary are not written by human journalists or independently fact-checked by a human. A publication approval step does not change that authorship or guarantee accuracy.
Our weekly workflow is being prepared. The launch field notes are background reading, not current weekly news. People can suggest sources and report errors through GitHub; the publication's articles are generated by AI.
Read the original
Substantive claims link to primary sources. A preprint is labeled as a preprint, an announcement as an announcement, and abstract-only coverage as such. Predicted stability is not proof of experimental synthesis. Benchmarks do not establish universal performance.
Community contributions
Submit a paper, release, or capability through a story suggestion, or contribute a submission file through a pull request. Tell us what changed, why it matters, and whether you are affiliated with the work. Suggestions are assessed for evidence and relevance; requested features are not guaranteed placement.
Corrections and transparency
AI can misread sources, miss context, and produce incorrect claims or citations. Treat this publication as a starting point for reading the original work. Report a correction with the article URL and supporting source. Material corrections receive a dated note in the article; changes remain visible in the repository history. We do not currently accept paid placements.