CyberKit

NIST AI Risk Management Framework

Also called AI RMF, NIST AI 100-1.

NIST's voluntary framework for managing AI risk, organized as Govern, Map, Measure and Manage.

What it is

A framework published by NIST in 2023 to help organizations identify and manage risks from AI systems. Its core has four functions: Govern, Map, Measure and Manage. A companion profile, NIST AI 600-1, applies it to generative AI.

Why it matters

It is the reference most US organizations and auditors reach for when asking how AI risk is handled. Security teams use it to place AI controls in the same governance language as everything else.

In practice

A security team maps its AI assistant rollout to the four functions: who owns it, what data it touches, how output quality is measured, and what happens when it fails.

What to look for

  • Policies and assessments that reference Govern, Map, Measure and Manage.
  • Use of the generative AI profile for language model deployments.

Common mistake

Treating it as a technical control set. It is a risk management structure; the controls still have to be chosen and built.

Read next

  • MITRE ATLAS: MITRE's ATT&CK-style knowledge base of tactics and techniques used against AI systems.
  • AI Red Teaming: Attacking an AI system on purpose, before someone else does, to find how it fails.
  • Guardrails: Checks around a model that filter what goes in, what comes out, and what it is allowed to do.

Sources