Open-weight artificial-intelligence models — those with publicly available parameters — are becoming essential scientific infrastructure (Y. Gal and S. Casper Nature 646, 286–287; 2025). They broaden access, accelerate replication and enable independent testing. But they also pose a risk: once downloaded, parameters can be modified and safeguards removed. This makes it harder to prevent and trace misuse than in monitored, hosted systems. Access Nature and 54 other Nature Portfolio journals Get Nature+, our best-value online-access subscription $32.99 / 30 days cancel any time Subscribe to this journal Receive 51 print issues and online access $199.00 per year only $3.90 per issue Rent or buy this article Prices vary by article type from$1.95 to$39.95 Prices may be subject to local taxes which are calculated during checkout Nature 651, 274 (2026) doi: https://doi.org/10.1038/d41586-026-00679-6 The authors declare no competing interests.…