Risk Management Processes for AI: Sizing AI risk management to the use case

Auditors ask about risk management processes for AI early, because the answer tells them how the rest of the program is likely to look. The outcome is NIST AI RMF GOVERN 1.3.

What the framework says

Processes, procedures, and practices are in place to determine the needed level of risk management activities based on the organization’s risk tolerance.

That is the text of NIST AI RMF GOVERN 1.3. Read it closely: it describes an outcome to achieve, not a product to buy, which is why two practices of the same size can satisfy it in different ways and both be right.

What this looks like in practice

In an assessment, NIST AI RMF GOVERN 1.3 is not one question. It resolves into several, and each one is really asking for different evidence. These are among the questions asked about it:

• Does the organization have a documented process that sets the depth of risk management for each AI system, review, testing, approval and monitoring, according to its approved AI risk tolerance?
• Do the organization's AI risk-tiering criteria account for generative AI factors, such as potential for malicious use, psychological impacts like emotional entanglement, obscene or untruthful output and dependencies on other systems?
• Has the organization set criteria for judging which AI risks it will accept and which it will act on, including the thresholds for treatment and the authority needed to accept a risk?

Ownership usually sits with the AI governance lead, though the evidence is often produced by someone else, which is where the trail breaks. A control that works but has no owner tends to stop working the month the person who quietly maintained it changes roles.

Who this applies to

NIST AI RMF is voluntary rather than binding, and that changes how it should be used. Adopting it does not create a legal obligation, and declining it is not a violation. It earns its place when a customer contract names it, when it gives structure to AI governance work you already owe under a law or a contract, or when a recognized practice is worth having on record. Treat it as a way of organizing that work, not as a second regime.

Building it, in order

The sequence below is the order that produces evidence as a by-product rather than as a separate documentation exercise:

1. Reference the approved AI risk tolerance statement.
2. Define risk tiers (for example low, moderate, high) with criteria for assigning an AI system to each tier.
3. Specify the risk management activities required at each tier, review depth, testing, approvals and monitoring frequency.
4. Review the organization's risk-tiering criteria against the generative AI factors NIST lists, information integrity, dependencies on other systems, harm to rights or safety, harmful or untruthful output, psychological impacts, malicious use, new security vulnerabilities, uneven impact on groups and unreliable performance.
5. Add criteria for the factors the tiering does not yet weigh.
6. Re-tier existing generative AI systems against the updated criteria.

Where else this maps

The questions behind NIST AI RMF GOVERN 1.3 also map to NIST AI 600-1 2.1, 2.11, 2.12 and ISO/IEC 42001 6.1.1, which are voluntary guidance and standards rather than law; following them does not by itself establish compliance with any law. That matters for scoping: the same evidence can be written once and referenced from each, rather than rewritten per framework.

Evidence an assessor asks for

For this outcome the evidence types that satisfy it are policy and procedure. The distinction matters more than it looks: a policy states what you intend to do, a procedure states how, and a record proves it happened on a date. Auditors ask for all three, and a practice that has written the first two often has nothing for the third.

Date every document and keep the superseded versions. The framework sets no retention period, so this is our recommendation rather than a requirement: keep each version for as long as the system it describes is in use and for a period after it is retired, because a question about a past decision is answered from the version that was in force when it was made.

Where this usually goes wrong

In our question bank this is weighted Critical: a gap here is raised as a critical finding (a partial answer as high). The usual cause of a gap is drift: the control was implemented once, the environment changed, and nothing re-checked it. A dated review on a fixed cadence is cheaper than the remediation.

What to do about it

• Produce the evidence, not the intention. For NIST AI RMF GOVERN 1.3 that means a signed, dated policy with a named owner and a review date and a step-by-step procedure someone unfamiliar with the task could follow. If the only evidence is that someone knows how this works, you have a practice and not a record.
• Put a name on it. Your AI governance lead should be recorded as the owner. A control with no owner stops working the month the person who quietly maintained it changes roles, and nobody notices.
• Review it on a fixed date, not when you remember. In our question bank this is weighted Critical: a gap here is raised as a critical finding (a partial answer as high). Put the review on the calendar with an owner, and keep the dated output of each one.
• Do not treat this as a legal obligation. It is a voluntary framework. Adopt it where a customer contract or a regulator's guidance names it, or where it gives structure to AI governance work you owe for other reasons, and say in your own documentation which of those reasons applies.

The recurring lesson with risk management processes for AI is that documentation is not the paperwork after the control. It is the part an assessor can read, and therefore the part that has to exist.

Related outcomes

• Trustworthy AI Characteristics: Defining what trustworthy means for clinical AI
• Ongoing AI Monitoring: Monitoring a model after it goes live
• Legal and Regulatory Requirements for AI: Classifying an AI tool before deployment

References

• GOVERN 1.3, in NIST AI Risk Management Framework 1.0 (NIST AI 100-1). This is a voluntary framework, not a legal requirement. The quoted statement is NIST's own, from the Core tables of NIST AI 100-1.
• Crosswalk references in this post are the mappings in our question bank, and are not verified against the published text of those standards.
• This post describes what a framework outcome asks for and what evidence documents it. It is not legal advice, and your counsel owns the conclusion. State law may impose additional obligations of its own.