Section 2.2.6(d) of the AI Safety Governance Framework 3.0 lists, among real-world safety risks, a risk of loss of control over knowledge of nuclear, biological, chemical and missile weapons, stating that AI substantially lowers the threshold for obtaining specialist knowledge in such high-risk fields and that, with retrieval-augmented generation and without effective control, it could be maliciously used by criminals, extremist forces or terrorists.

One recorded claim. The claim panel gives its qualification, the source passages with their place in each source, and how it was checked.