NIST AI Risk Management Framework (AI RMF)

What the AI RMF is

The NIST AI Risk Management Framework is a free, voluntary framework from the US National Institute of Standards and Technology. Version 1.0 was released on 26 January 2023 to help organisations build trustworthiness into how they design, develop, use and evaluate AI systems. NIST has said the framework is being revised, so check the official page for the current state before you quote it.

It is organised around four functions:

• Govern: ownership, policies, roles and culture for AI risk.
• Map: understanding the context of a system, what it is for, who it affects and what could go wrong.
• Measure: assessing, analysing and tracking the risks you mapped.
• Manage: deciding how to treat each risk and acting on it.

On 26 July 2024 NIST added the Generative AI Profile (NIST AI 600-1), which applies the framework to risks that are specific to or made worse by generative AI. A companion Playbook suggests concrete actions for each part of the framework.

Is it worth your time?

For anyone on an AI security path, yes, at the level of understanding its structure and language. Much of real AI security work ends up as evidence for decisions: which risks were found, how they were measured and who accepted what. The AI RMF gives security, legal and product teams one shared way to talk about that. Red teaming and security evals sit naturally under Measure, while guardrails and incident plans sit under Manage.

If you are just starting, read it after you understand how LLM apps work and fail. Without that grounding it can feel abstract.

How long does it take?

Reading the core document and the Generative AI Profile summary takes a few hours. You do not need to memorise it. Return to it when you write a risk assessment or explain your testing to people outside engineering.

Prerequisites

None formally. A basic grasp of security risk (likelihood and impact) and of how LLM applications are built makes it far easier to read.

Free or paid?

Free. All documents are published by NIST on its website and in the AI Resource Center.

Common questions

Is it a law? No. It is voluntary guidance, although customers and regulators increasingly refer to it.

Is it a technical checklist? Not really. Much of it is about accountability, context and decisions. Pair it with the OWASP Top 10 for LLM Applications and MITRE ATLAS for the technical side.

How does it relate to the NIST Cybersecurity Framework? They are separate frameworks with a similar spirit: the CSF covers cybersecurity broadly, the AI RMF covers AI risk, including risks that are not about security at all, such as fairness.