Cloud Security Alliance, 2026-05-20; retrieved 2026-09-14 tier 4 press (primary)unaudited

CAISI Frontier Testing Agreements Reach Five Labs

DeepMind, Microsoft, xAI added May 2026; 40+ evaluations; TRAINS taskforce; 2025 refocus. Cited from a search snippet; not re-fetched in full.

US CAISI (NIST), F (for): Public funding; formal access agreements with five developers.“CAISI, which operates within NIST at the Department of Commerce”
US CAISI (NIST), A (for): More than 40 evaluations including unreleased models; classified CBRN and cyber work via the TRAINS taskforce.“completed more than 40 evaluations — including assessments of frontier models not yet available to the public”
US CAISI (NIST), G (against): Refocused in 2025 from broad safety research to demonstrable national-security risks; mandate is politically steerable.“repositioned the center as CAISI, shifting emphasis toward national security and cybersecurity risk reduction”
US CAISI (NIST), R (against): Publishes little per model; findings mostly stay inside government.“voluntary agreements, unclassified evaluations, no mandatory disclosure of findings”
Microsoft AI Red Team, F (against): Microsoft is a major OpenAI investor and a frontier developer with its own CAISI agreement.“Google (DeepMind), Microsoft, and xAI signed agreements with the US Center for AI Standards and Innovation (CAISI)”