AI risk management framework reference
NIST AI Risk Management Framework: functions, profiles and practice
What the NIST AI RMF 1.0 says, in NIST's own terms: the four functions and their categories, the characteristics of trustworthy AI, the Generative AI Profile, the Playbook, and how a regulated mid-market company can put it to work.
The NIST AI Risk Management Framework (AI RMF 1.0, published as NIST AI 100-1) is guidance from the National Institute of Standards and Technology, intended for voluntary use, for managing risks AI systems pose to individuals, organizations and society. It organizes that work into four functions, Govern, Map, Measure and Manage.
what it is
What the NIST AI Risk Management Framework is
NIST released AI RMF 1.0 on January 26, 2023, after a request for information, two public drafts and workshops. NIST's page says it 'is intended for voluntary use and to improve the ability to incorporate trustworthiness considerations into the design, development, use, and evaluation of AI products, services, and systems.'
NIST calls the framework 'voluntary, rights-preserving, non-sector-specific, and use-case agnostic'. It is not a regulation, and the Core's actions 'do not constitute a checklist'. NIST says it 'should be useful for those who design, develop, use, or evaluate AI technologies.' It is not for anyone wanting a certificate or a sign-off checklist.
A status note for anyone writing policy against it: NIST's page states 'The AI RMF 1.0 is being revised as part of the White House AI Action Plan.' Build to outcomes, not clause numbers.
four functions
The four functions of the AI RMF: Govern, Map, Measure and Manage
In NIST AI 100-1, Govern cultivates a culture of risk management and is 'a cross-cutting function' across the other three. Map 'establishes the context to frame risks related to an AI system'. Measure uses quantitative, qualitative or mixed methods 'to analyze, assess, benchmark, and monitor AI risk and related impacts'. Manage allocates resources to mapped and measured risks, including incident response and recovery. The nineteen categories, in NIST's words:
NIST AI governance: why Govern comes first
Inside the framework, NIST AI governance is the GOVERN function. NIST AI 100-1 says that 'after instituting the outcomes in GOVERN, most users of the AI RMF would start with the MAP function and continue to MEASURE or MANAGE', and that the process should be iterative.
| Category | Outcome, in NIST's words |
|---|---|
| GOVERN 1 | Policies, processes, procedures, and practices across the organization related to the mapping, measuring, and managing of AI risks are in place, transparent, and implemented effectively. |
| GOVERN 2 | Accountability structures are in place so that the appropriate teams and individuals are empowered, responsible, and trained for mapping, measuring, and managing AI risks. |
| GOVERN 3 | Workforce diversity, equity, inclusion, and accessibility processes are prioritized in the mapping, measuring, and managing of AI risks throughout the lifecycle. |
| GOVERN 4 | Organizational teams are committed to a culture that considers and communicates AI risk. |
| GOVERN 5 | Processes are in place for robust engagement with relevant AI actors. |
| GOVERN 6 | Policies and procedures are in place to address AI risks and benefits arising from third-party software and data and other supply chain issues. |
| MAP 1 | Context is established and understood. |
| MAP 2 | Categorization of the AI system is performed. |
| MAP 3 | AI capabilities, targeted usage, goals, and expected benefits and costs compared with appropriate benchmarks are understood. |
| MAP 4 | Risks and benefits are mapped for all components of the AI system including third-party software and data. |
| MAP 5 | Impacts to individuals, groups, communities, organizations, and society are characterized. |
| MEASURE 1 | Appropriate methods and metrics are identified and applied. |
| MEASURE 2 | AI systems are evaluated for trustworthy characteristics. |
| MEASURE 3 | Mechanisms for tracking identified AI risks over time are in place. |
| MEASURE 4 | Feedback about efficacy of measurement is gathered and assessed. |
| MANAGE 1 | AI risks based on assessments and other analytical output from the MAP and MEASURE functions are prioritized, responded to, and managed. |
| MANAGE 2 | Strategies to maximize AI benefits and minimize negative impacts are planned, prepared, implemented, documented, and informed by input from relevant AI actors. |
| MANAGE 3 | AI risks and benefits from third-party entities are managed. |
| MANAGE 4 | Risk treatments, including response and recovery, and communication plans for the identified and measured AI risks are documented and monitored regularly. |
trustworthy ai
The characteristics of trustworthy AI
NIST AI 100-1 lists them as: valid and reliable, safe, secure and resilient, accountable and transparent, explainable and interpretable, privacy-enhanced, and fair with harmful bias managed.
Its Figure 4 shows valid and reliable as 'a necessary condition of trustworthiness' and accountable and transparent as relating to all the others. NIST warns that addressing the characteristics individually 'will not ensure AI system trustworthiness', that tradeoffs are involved, and that human judgment should set the specific metrics and thresholds.
profiles
AI RMF profiles and the Generative AI Profile
A profile applies the functions, categories and subcategories to a specific setting. NIST AI 100-1 describes three kinds and 'does not prescribe profile templates'.
NIST AI 600-1: the Generative AI Profile
Released July 26, 2024, NIST AI 600-1 is 'a cross-sectoral profile of and companion resource for the AI Risk Management Framework (AI RMF 1.0) for Generative AI'. It defines twelve risks unique to or exacerbated by generative AI: CBRN information or capabilities; confabulation; dangerous, violent, or hateful content; data privacy; environmental impacts; harmful bias or homogenization; human-AI configuration; information integrity; information security; intellectual property; obscene, degrading, and/or abusive content; and value chain and component integration. Confabulation is 'the production of confidently stated but erroneous or false content', known colloquially as hallucinations. Its suggested actions focus on Governance, Content Provenance, Pre-deployment Testing and Incident Disclosure.
On April 7, 2026, NIST also released a concept note for an AI RMF Profile on Trustworthy AI in Critical Infrastructure.
1. Use-case profiles
The framework applied to one setting or application, such as hiring.
2. Temporal profiles
A Current Profile describes how AI risk is managed now; a Target Profile describes the outcomes needed. NIST notes that comparing them 'likely reveals gaps' for action plans to address.
3. Cross-sectoral profiles
Risks of models or applications used across sectors, such as large language models.
playbook
The NIST AI risk management framework Playbook
The AI RMF Playbook 'provides suggested actions for achieving the outcomes laid out in the AI Risk Management Framework (AI RMF) Core', aligned to each subcategory.
NIST says it 'is neither a checklist nor set of steps to be followed in its entirety'; organizations borrow the suggestions that apply. It downloads as PDF, CSV, Excel or JSON. NIST states it 'will be updated after the AI RMF 1.0 is revised', so record which suggestions you adopted.
in practice
Putting an AI risk management framework into practice at a regulated mid-market company
This sequence is our practice, not a NIST requirement; each step names the categories it serves.
AI model governance, AI risk governance and technical AI governance
AI model governance controls individual systems (inventory, testing, change control, monitoring). AI risk governance is the organizational layer in Govern. Technical AI governance is the engineering that enforces both, such as logging and approval checkpoints.
1. Owner and policy (GOVERN 1, GOVERN 2)
Name who is accountable for AI risk, who approves a new AI use and who can pause one, and write it down. In RMF terms, a NIST AI policy starts here, before any tool.
2. AI inventory (MAP 1, MAP 4)
List every AI system in production or pilot, including AI inside vendor software: purpose, users, data, affected people and third-party components.
3. Risk tiers (MAP 2, MAP 5, MANAGE 1)
Tiering is not NIST's term; it is how an organization turns categorization and impact into priorities, so review effort follows risk.
4. Measures per tier (MEASURE 1, MEASURE 2)
Decide which tests each tier needs before release and in operation, and who sets the thresholds.
5. Third parties (GOVERN 6, MANAGE 3)
Record what you know about each vendor's model and data, what you cannot see, and how that gap is handled.
6. Records and profiles (MEASURE 3, MANAGE 4)
Keep each review's record, set monitoring and an incident path, and compare Current and Target Profiles.
beyond the rmf
AI risk management beyond the RMF: ISO/IEC 42001 and the EU AI Act
ISO/IEC 42001
ISO says ISO/IEC 42001 'specifies requirements for establishing, implementing, maintaining, and continually improving an Artificial Intelligence Management System (AIMS)'. The RMF describes risk outcomes; the ISO standard sets requirements for the management system around them.
The EU AI Act
The European Commission describes the AI Act, Regulation (EU) 2024/1689, as a legal framework setting risk-based rules for AI developers and deployers, with four levels of risk: unacceptable, high, transparency, and minimal or no risk. Unlike the RMF it is law; whether it applies to a US company is a question for counsel.
A note for banks: SR 26-2
SR 26-2 states that generative AI and agentic AI models are not within its scope, and that it does not set forth enforceable standards. Your own risk management should still govern them. See our SR 26-2 model risk page and our financial services AI governance page.
where kriv fits
Where Kriv AI fits
Kriv AI works with regulated mid-market companies in healthcare, life sciences, financial services and insurance.
We write the rules to the policy and frameworks you already answer to, such as the NIST AI RMF, ISO/IEC 42001, the EU AI Act, HIPAA or the NAIC Model Bulletin. We do not certify compliance and make no claim of formal alignment with the RMF. What we build for each function is on our AI governance consulting page; starting from a blank page, see our AI governance framework guide.
Kriv AI sends a senior practitioner who both scopes and delivers: no handoffs, no bench, no discovery theater. Current and Target Profiles or a category-by-category gap list can be prepared on request. The honest gap: we have no named NIST AI RMF case study we can publish.
get started
Get started
Work is scoped and priced after a short discovery call. Book an AI RMF discovery call, or see how work is scoped and priced. Governance versus compliance? Read our explainer on the difference.
Sources
Cited sources
- NIST, AI Risk Management Framework landing page (read 2026-10-20)
- NIST AI 100-1, Artificial Intelligence Risk Management Framework (AI RMF 1.0), January 2023 (read 2026-10-20)
- NIST AI 600-1, AI RMF: Generative Artificial Intelligence Profile, July 2024 (read 2026-10-20)
- NIST AI RMF Playbook, Trustworthy and Responsible AI Resource Center (read 2026-10-20)
- NIST, NIST AI RMF Playbook page, updated June 10, 2026 (read 2026-10-20)
- NIST, AI Risk Management Framework FAQs, updated August 13, 2026 (read 2026-10-20)
- NIST, AI RMF Development, updated March 27, 2026 (read 2026-10-20)
- NIST, Crosswalks to the AI RMF 1.0, updated December 17, 2024 (read 2026-10-20)
- NIST, Concept Note: AI RMF Profile on Trustworthy AI in Critical Infrastructure, updated July 17, 2026 (read 2026-10-20)
- ISO, ISO/IEC 42001:2023 Artificial intelligence management system (read 2026-10-20)
- European Commission, AI Act regulatory framework page (read 2026-10-20)
Straight answers
Frequently asked questions about NIST AI Risk Management Framework: functions, profiles and practice
Is the NIST AI Risk Management Framework mandatory?
NIST says no. Its FAQ asks whether organizations will be required to use the framework and answers: 'No. NIST has produced the AI RMF as a voluntary Framework.' Whether anything else you answer to refers to it is a question for your compliance and legal teams.
What are the four functions of the NIST AI RMF?
Govern, Map, Measure and Manage. Govern sets policy and accountability and runs across the other three. Map establishes each AI system's context. Measure analyzes, assesses and monitors AI risk. Manage allocates resources to the mapped and measured risks, including response and recovery.
What is the NIST AI RMF Playbook?
NIST's companion resource of suggested actions for each subcategory of the four functions. NIST says it 'is neither a checklist nor set of steps to be followed in its entirety'. It downloads as PDF, CSV, Excel or JSON, and NIST will update it after the AI RMF 1.0 is revised.
What is NIST AI 600-1, the Generative AI Profile?
Released by NIST on July 26, 2024, it is a cross-sectoral profile of the AI RMF 1.0 for generative AI. It defines twelve risks unique to or exacerbated by generative AI, including confabulation, data privacy and information security, and suggests actions to govern, map, measure and manage them.
Is the NIST AI RMF 1.0 being revised?
Yes. NIST's page states that 'the AI RMF 1.0 is being revised as part of the White House AI Action Plan', and its FAQ says the plan, dated July 23, 2025, tasked NIST with the revision. NIST has not said there what will change or when.
What does NIST AI governance mean in the AI RMF?
It is the GOVERN function. Its six categories cover policies and procedures, accountability, workforce diversity and accessibility processes, a culture that communicates AI risk, engagement with AI actors, and third-party risk. NIST calls it cross-cutting: it informs Map, Measure and Manage.
How does the NIST AI RMF relate to ISO/IEC 42001?
They do different jobs. The AI RMF is voluntary guidance describing risk management outcomes in four functions. ISO/IEC 42001, per ISO, specifies requirements for an AI management system. A team using both can map one set of evidence to each rather than collecting it twice.
Where should a mid-market company start with the AI RMF?
With Govern, then Map, the order NIST AI 100-1 describes for most users. That means naming an accountable owner and approval path, then inventorying every AI system in use, including AI inside vendor software. Tiers, measures and monitoring follow from the inventory.
Talk to the team that would do the work
Bring your requirements to a working session with the person who'll actually deliver.
Book a Discovery Call