AI · Security · 2026-09-25 15:27
Declassified estimates reveal that the NSA's AI safety center is spending enormous sums testing frontier models to identify national security vulnerabilities, far exceeding previously known costs.
Auto-generated summary for reference only. RSS