Achieving AI Model Weight Security Level 3 (SL3)
Source: Rand.org· Jair Aguirre, Phillip Robertson, Steven F. Comer, Matthew J. Malone, Wesley Hurd· August 25, 2026
RAND proposes Security Level 3 (SL3), a framework of 262 security controls to protect artificial intelligence (AI) model weights against organized cybercrime and insider threats, achievable within six to 12 months.
This story was reported by Rand.org. Read the full original article:
Read on Rand.org 
