NIST renames its AI safety consortium
On 29 May 2026 NIST published notice 2026-10779 in the Federal Register: the Artificial Intelligence Safety Institute Consortium (AISIC), formed on 2 November 2023, is retitled the NIST Artificial Intelligence Consortium, its research scope is revised, and organisations are again invited to submit letters of interest.
Why it matters
The state body that sets standards for measuring AI changed the name and the list of tasks of its joint venue with industry, and the word 'safety' left the name. An editorial assessment: the notice only renames and rewrites the scope; it does not say what became of the earlier work.
What it says. The notice recalls AISIC, which it says brought together more than 280 organisations, and names nine areas of the new plan: free speech and American values in frontier systems, open-source and open-weight AI, adoption, AI-enabled science, interpretability, control and robustness, an evaluations ecosystem, protection of commercial and government innovations, secure-by-design AI, and the government's place in evaluating national-security risks of frontier models. Participants sign a CRADA; letters are accepted continuously with reviews likely twice a year, the first within 60 days of publication; current members need not reapply. What the record does not claim. The notice says nothing about the fate of documents produced under AISIC. The figure 'more than 280' is NIST's own statement about 2023 and was not checked here.