NIST AI Risk Management Framework

United States

5 domains

284 controls

NIST AI 100-1 AI RMF Core + AI 600-1 Generative AI Profile. Voluntary AI risk management framework.

Verified

Framework summaries on this platform are AI-assisted interpretations for educational and compliance planning purposes. They do not reproduce or replace the official standards. Refer to the authoritative source for the definitive text. Framework names and trademarks belong to their respective organisations.

Framework Domains (5)

Generative AI Profile

212 controls

Controls in the Generative AI Profile domain of NIST AI Risk Management Framework — 212 controls
Code	Title	Description
GAI-GV-1.1-001	GAI Profile action GV-1.1-001 mapped to GOVERN-1.1	Align GAI policies with applicable laws and regulations including privacy, IP, consumer protection, anti-discrimination,...
GAI-GV-1.1-002	GAI Profile action GV-1.1-002 mapped to GOVERN-1.1	Track and integrate GAI specific regulatory developments and updates such as guidance, directives, and case law into the...
GAI-GV-1.2-001	GAI Profile action GV-1.2-001 mapped to GOVERN-1.2	Codify trustworthy GAI characteristics including safety, security, fairness, transparency, accountability, explainabilit...
GAI-GV-1.2-002	GAI Profile action GV-1.2-002 mapped to GOVERN-1.2	Establish documented safety and risk reduction measures applied both prior to deployment and on an ongoing basis, includ...
GAI-GV-1.3-001	GAI Profile action GV-1.3-001 mapped to GOVERN-1.3	Determine the level of GAI risk management activities based on use case risk tier, capability, and downstream exposure.
GAI-GV-1.3-002	GAI Profile action GV-1.3-002 mapped to GOVERN-1.3	Periodically evaluate GAI systems for risks linked to confabulation and dangerous, violent, or hateful content.
GAI-GV-1.3-003	GAI Profile action GV-1.3-003 mapped to GOVERN-1.3	Periodically evaluate whether GAI models may misuse CBRN information or otherwise provide uplift to high consequence mis...
GAI-GV-1.3-004	GAI Profile action GV-1.3-004 mapped to GOVERN-1.3	Calibrate controls for obscene, degrading, or abusive content and harmful bias risks at appropriate review frequency.
GAI-GV-1.3-005	GAI Profile action GV-1.3-005 mapped to GOVERN-1.3	Establish review frequency for harmful bias and homogenization risks based on training data composition and downstream c...
GAI-GV-1.3-006	GAI Profile action GV-1.3-006 mapped to GOVERN-1.3	Set policies covering violent or hateful content and CBRN risks across the GAI lifecycle from concept to retirement.
GAI-GV-1.3-007	GAI Profile action GV-1.3-007 mapped to GOVERN-1.3	Set policies governing information security and information integrity risks including model abuse, prompt injection, and...
GAI-GV-1.4-001	GAI Profile action GV-1.4-001 mapped to GOVERN-1.4	Establish transparent policies and procedures addressing harmful bias and homogenization across GAI use cases.
GAI-GV-1.4-002	GAI Profile action GV-1.4-002 mapped to GOVERN-1.4	Establish transparent policies and procedures addressing obscene, degrading, or abusive content generated by GAI systems...
GAI-GV-1.5-001	GAI Profile action GV-1.5-001 mapped to GOVERN-1.5	Plan ongoing monitoring and periodic review of GAI risk management for information integrity outcomes.
GAI-GV-1.5-002	GAI Profile action GV-1.5-002 mapped to GOVERN-1.5	Conduct after action reviews of GAI system incidents and incident disclosures to identify gaps and update controls.
GAI-GV-1.5-003	GAI Profile action GV-1.5-003 mapped to GOVERN-1.5	Refresh test, evaluation, validation, and verification (TEVV) and digital content transparency methods for GAI on a defi...
GAI-GV-1.6-001	GAI Profile action GV-1.6-001 mapped to GOVERN-1.6	Maintain a GAI inventory including model lineage, training data sources, intended use cases, and deployment owners.
GAI-GV-1.6-002	GAI Profile action GV-1.6-002 mapped to GOVERN-1.6	Include GAI system risk classifications and deployment context (e.g., consumer-facing, internal-only) in the inventory.
GAI-GV-1.6-003	GAI Profile action GV-1.6-003 mapped to GOVERN-1.6	Resource GAI inventory maintenance according to organizational risk priorities, with ownership in the AI governance func...
GAI-GV-1.7-001	GAI Profile action GV-1.7-001 mapped to GOVERN-1.7	Establish processes for decommissioning GAI systems safely without increasing residual risk or eroding stakeholder trust...
GAI-GV-1.7-002	GAI Profile action GV-1.7-002 mapped to GOVERN-1.7	Plan secure data retention, deletion, or transfer for retired GAI models and associated training, fine-tuning, and promp...
GAI-GV-2.1-001	GAI Profile action GV-2.1-001 mapped to GOVERN-2.1	Document GAI roles, responsibilities, and communication lines covering developers, deployers, operators, governance, leg...
GAI-GV-2.1-002	GAI Profile action GV-2.1-002 mapped to GOVERN-2.1	Define and document escalation paths for GAI incidents, risk acceptance decisions, and policy exceptions.
GAI-GV-2.1-003	GAI Profile action GV-2.1-003 mapped to GOVERN-2.1	Train personnel on GAI roles and responsibilities including red teaming, content provenance, and incident response.
GAI-GV-2.1-004	GAI Profile action GV-2.1-004 mapped to GOVERN-2.1	Document responsibilities for managing pre-trained or foundation model dependencies, including update and re-evaluation...
GAI-GV-2.1-005	GAI Profile action GV-2.1-005 mapped to GOVERN-2.1	Document responsibilities for content provenance, watermarking, and disclosure of AI-generated output.
GAI-GV-3.2-001	GAI Profile action GV-3.2-001 mapped to GOVERN-3.2	Define and differentiate roles for human-GAI configurations, distinguishing oversight, review, and override responsibili...
GAI-GV-3.2-002	GAI Profile action GV-3.2-002 mapped to GOVERN-3.2	Address risks of over-reliance and automation bias in human-GAI configurations through training and UX design.
GAI-GV-3.2-003	GAI Profile action GV-3.2-003 mapped to GOVERN-3.2	Set policies for safe, ethical, and accountable use of GAI by employees including acceptable use limitations.
GAI-GV-3.2-004	GAI Profile action GV-3.2-004 mapped to GOVERN-3.2	Verify that GAI human oversight roles have authority, training, and time to meaningfully review outputs.
GAI-GV-3.2-005	GAI Profile action GV-3.2-005 mapped to GOVERN-3.2	Establish governance for GAI use in sensitive contexts such as hiring, credit, healthcare, and law enforcement.
GAI-GV-4.1-001	GAI Profile action GV-4.1-001 mapped to GOVERN-4.1	Establish policies fostering critical thinking and a safety-first mindset in GAI design, development, and use.
GAI-GV-4.1-002	GAI Profile action GV-4.1-002 mapped to GOVERN-4.1	Establish whistleblower or speak-up channels for raising concerns about GAI risks without retaliation.
GAI-GV-4.1-003	GAI Profile action GV-4.1-003 mapped to GOVERN-4.1	Document organizational risk tolerance for GAI specific harms including confabulation, bias, and misuse.
GAI-GV-4.2-001	GAI Profile action GV-4.2-001 mapped to GOVERN-4.2	Document GAI risks and potential impacts including model cards and system cards for deployed systems.
GAI-GV-4.2-002	GAI Profile action GV-4.2-002 mapped to GOVERN-4.2	Communicate GAI risks more broadly through transparency reports, disclosures, and stakeholder briefings.
GAI-GV-4.2-003	GAI Profile action GV-4.2-003 mapped to GOVERN-4.2	Document risks and impacts in machine readable formats to enable downstream consumption and integration.
GAI-GV-4.3-002	GAI Profile action GV-4.3-002 mapped to GOVERN-4.3	Enable GAI testing including red teaming, adversarial evaluation, and safety evaluation aligned to deployment context.
GAI-GV-4.3-003	GAI Profile action GV-4.3-003 mapped to GOVERN-4.3	Participate in incident information sharing forums and disclose GAI vulnerabilities responsibly.
GAI-GV-5.1-001	GAI Profile action GV-5.1-001 mapped to GOVERN-5.1	Engage external stakeholders for feedback on GAI societal impacts including affected communities.
GAI-GV-5.1-002	GAI Profile action GV-5.1-002 mapped to GOVERN-5.1	Integrate adjudicated external feedback into GAI design, deployment, and policy updates.
GAI-GV-6.1-001	GAI Profile action GV-6.1-001 mapped to GOVERN-6.1	Address GAI risks from third-party entities including data, model, and service providers in policy.
GAI-GV-6.1-002	GAI Profile action GV-6.1-002 mapped to GOVERN-6.1	Document third-party GAI dependencies and require disclosure of training data sources, license terms, and known limitati...
GAI-GV-6.1-003	GAI Profile action GV-6.1-003 mapped to GOVERN-6.1	Address intellectual property and licensing risks from third-party GAI components in contracts and policy.
GAI-GV-6.1-004	GAI Profile action GV-6.1-004 mapped to GOVERN-6.1	Require third-party GAI providers to disclose safety evaluations, incident history, and security posture.
GAI-GV-6.1-005	GAI Profile action GV-6.1-005 mapped to GOVERN-6.1	Address data residency and cross-border transfer risks for third-party GAI usage.
GAI-GV-6.1-006	GAI Profile action GV-6.1-006 mapped to GOVERN-6.1	Address risks of using third-party GAI in regulated workloads through tailored due diligence.
GAI-GV-6.1-007	GAI Profile action GV-6.1-007 mapped to GOVERN-6.1	Document acceptable use restrictions imposed by third-party GAI providers and reflect them downstream.
GAI-GV-6.1-008	GAI Profile action GV-6.1-008 mapped to GOVERN-6.1	Define service level expectations including model deprecation and notification timelines for third-party GAI.
GAI-GV-6.1-009	GAI Profile action GV-6.1-009 mapped to GOVERN-6.1	Reserve audit, evaluation, or attestation rights for high risk third-party GAI components.
GAI-GV-6.1-010	GAI Profile action GV-6.1-010 mapped to GOVERN-6.1	Plan for third-party GAI license changes, price changes, or termination through alternate providers.
GAI-GV-6.2-001	GAI Profile action GV-6.2-001 mapped to GOVERN-6.2	Establish contingency processes for failures or incidents in third-party GAI data or systems deemed high risk.
GAI-GV-6.2-002	GAI Profile action GV-6.2-002 mapped to GOVERN-6.2	Maintain fallback procedures including human review or alternate models when a third-party GAI service degrades.
GAI-GV-6.2-003	GAI Profile action GV-6.2-003 mapped to GOVERN-6.2	Document data and prompt portability so workloads can be moved between third-party GAI providers.
GAI-GV-6.2-004	GAI Profile action GV-6.2-004 mapped to GOVERN-6.2	Maintain incident response runbooks for third-party GAI outages, data leakage, or output quality issues.
GAI-GV-6.2-005	GAI Profile action GV-6.2-005 mapped to GOVERN-6.2	Coordinate with third-party GAI providers on shared responsibility for incident response and disclosure.
GAI-GV-6.2-006	GAI Profile action GV-6.2-006 mapped to GOVERN-6.2	Test third-party GAI failover procedures periodically and document results.
GAI-GV-6.2-007	GAI Profile action GV-6.2-007 mapped to GOVERN-6.2	Communicate third-party GAI incidents to affected internal and external stakeholders in a timely manner.
GAI-MG-1.3-001	GAI Profile action MG-1.3-001 mapped to MANAGE-1.3	Develop, plan, and document responses to high priority GAI risks including mitigating, transferring, avoiding, or accept...
GAI-MG-1.3-002	GAI Profile action MG-1.3-002 mapped to MANAGE-1.3	Document residual GAI risk acceptance with sign off from accountable executives.
GAI-MG-2.2-001	GAI Profile action MG-2.2-001 mapped to MANAGE-2.2	Implement mechanisms to sustain GAI system value including monitoring of effectiveness, accuracy, and user satisfaction.
GAI-MG-2.2-002	GAI Profile action MG-2.2-002 mapped to MANAGE-2.2	Implement content provenance mechanisms such as watermarking and detection tools where appropriate.
GAI-MG-2.2-003	GAI Profile action MG-2.2-003 mapped to MANAGE-2.2	Implement controls for safe and responsible GAI use including content filters and refusal policies.
GAI-MG-2.2-004	GAI Profile action MG-2.2-004 mapped to MANAGE-2.2	Implement controls for human-AI configuration including required review of high consequence outputs.
GAI-MG-2.2-005	GAI Profile action MG-2.2-005 mapped to MANAGE-2.2	Implement controls for confabulation including factuality checks and citation verification.
GAI-MG-2.2-006	GAI Profile action MG-2.2-006 mapped to MANAGE-2.2	Implement controls for harmful bias including ongoing disaggregated performance monitoring.
GAI-MG-2.2-007	GAI Profile action MG-2.2-007 mapped to MANAGE-2.2	Implement controls for information security including prompt injection defenses and data leakage prevention.
GAI-MG-2.2-008	GAI Profile action MG-2.2-008 mapped to MANAGE-2.2	Implement controls for intellectual property including copyright detection and license enforcement.
GAI-MG-2.2-009	GAI Profile action MG-2.2-009 mapped to MANAGE-2.2	Implement controls for data privacy including PII detection and redaction in prompts and outputs.
GAI-MG-2.3-001	GAI Profile action MG-2.3-001 mapped to MANAGE-2.3	Respond to and recover from previously unknown GAI risks including emergent capabilities and novel attacks.
GAI-MG-2.4-001	GAI Profile action MG-2.4-001 mapped to MANAGE-2.4	Establish mechanisms to supersede, disengage, or deactivate GAI systems exhibiting unexpected behavior.
GAI-MG-2.4-002	GAI Profile action MG-2.4-002 mapped to MANAGE-2.4	Establish kill switch and rollback procedures for GAI models in production.
GAI-MG-2.4-003	GAI Profile action MG-2.4-003 mapped to MANAGE-2.4	Test GAI supersede, disengage, and deactivate procedures periodically.
GAI-MG-2.4-004	GAI Profile action MG-2.4-004 mapped to MANAGE-2.4	Document GAI deactivation decisions, residual risk, and stakeholder communication.
GAI-MG-3.1-001	GAI Profile action MG-3.1-001 mapped to MANAGE-3.1	Monitor GAI risks and benefits from third-party resources on an ongoing basis.
GAI-MG-3.1-002	GAI Profile action MG-3.1-002 mapped to MANAGE-3.1	Apply and document risk controls for third-party GAI components.
GAI-MG-3.1-003	GAI Profile action MG-3.1-003 mapped to MANAGE-3.1	Track third-party GAI provider security and safety advisories and incidents.
GAI-MG-3.1-004	GAI Profile action MG-3.1-004 mapped to MANAGE-3.1	Verify ongoing third-party GAI compliance with contractual safety and security expectations.
GAI-MG-3.1-005	GAI Profile action MG-3.1-005 mapped to MANAGE-3.1	Re-assess third-party GAI risk on a defined cadence and upon material change.
GAI-MG-3.2-001	GAI Profile action MG-3.2-001 mapped to MANAGE-3.2	Monitor pre-trained models used in GAI development as part of regular monitoring and maintenance.
GAI-MG-3.2-002	GAI Profile action MG-3.2-002 mapped to MANAGE-3.2	Track pre-trained GAI model version changes and deprecation announcements.
GAI-MG-3.2-003	GAI Profile action MG-3.2-003 mapped to MANAGE-3.2	Test pre-trained GAI models for safety and bias regression after upstream updates.
GAI-MG-3.2-004	GAI Profile action MG-3.2-004 mapped to MANAGE-3.2	Document pre-trained GAI model lineage, training data disclosures, and license terms.
GAI-MG-3.2-005	GAI Profile action MG-3.2-005 mapped to MANAGE-3.2	Apply controls to mitigate inherited risks from pre-trained GAI models.
GAI-MG-3.2-006	GAI Profile action MG-3.2-006 mapped to MANAGE-3.2	Re-evaluate fine-tuning impact on pre-trained GAI model behavior.
GAI-MG-3.2-007	GAI Profile action MG-3.2-007 mapped to MANAGE-3.2	Document pre-trained GAI model evaluation outcomes and remediation actions.
GAI-MG-3.2-008	GAI Profile action MG-3.2-008 mapped to MANAGE-3.2	Coordinate with pre-trained GAI providers on safety and security incident response.
GAI-MG-3.2-009	GAI Profile action MG-3.2-009 mapped to MANAGE-3.2	Plan alternatives to pre-trained GAI models in case of deprecation or unacceptable risk.
GAI-MG-4.1-001	GAI Profile action MG-4.1-001 mapped to MANAGE-4.1	Implement post-deployment GAI monitoring plans including drift, performance, and safety signals.
GAI-MG-4.1-002	GAI Profile action MG-4.1-002 mapped to MANAGE-4.1	Implement GAI user feedback capture and adjudication processes.
GAI-MG-4.1-003	GAI Profile action MG-4.1-003 mapped to MANAGE-4.1	Implement GAI appeal and override mechanisms with documented criteria.
GAI-MG-4.1-004	GAI Profile action MG-4.1-004 mapped to MANAGE-4.1	Implement GAI decommissioning and change management procedures.
GAI-MG-4.1-005	GAI Profile action MG-4.1-005 mapped to MANAGE-4.1	Implement GAI incident response procedures aligned to enterprise IR program.
GAI-MG-4.1-006	GAI Profile action MG-4.1-006 mapped to MANAGE-4.1	Implement GAI logging and observability to support investigation and audit.
GAI-MG-4.1-007	GAI Profile action MG-4.1-007 mapped to MANAGE-4.1	Test GAI monitoring procedures periodically including simulated incidents.
GAI-MG-4.2-001	GAI Profile action MG-4.2-001 mapped to MANAGE-4.2	Integrate measurable activities for GAI continual improvement into update cycles.
GAI-MG-4.2-002	GAI Profile action MG-4.2-002 mapped to MANAGE-4.2	Engage interested parties including AI actors and affected communities in GAI improvement.
GAI-MG-4.2-003	GAI Profile action MG-4.2-003 mapped to MANAGE-4.2	Document GAI improvement outcomes and route them back into governance, mapping, and measurement.
GAI-MG-4.3-001	GAI Profile action MG-4.3-001 mapped to MANAGE-4.3	Communicate GAI incidents and errors to relevant AI actors including affected communities.
GAI-MG-4.3-002	GAI Profile action MG-4.3-002 mapped to MANAGE-4.3	Document GAI incident tracking, response, and recovery actions.
GAI-MG-4.3-003	GAI Profile action MG-4.3-003 mapped to MANAGE-4.3	Share GAI incident insights externally where appropriate including industry information sharing organizations.
GAI-MP-1.1-001	GAI Profile action MP-1.1-001 mapped to MAP-1.1	Document intended purpose, beneficial uses, and prospective settings for the GAI system including users and impacted par...
GAI-MP-1.1-002	GAI Profile action MP-1.1-002 mapped to MAP-1.1	Document context specific laws, norms, and expectations applicable to the GAI use case.
GAI-MP-1.1-003	GAI Profile action MP-1.1-003 mapped to MAP-1.1	Document GAI system assumptions and limitations across the lifecycle and product roadmap.
GAI-MP-1.1-004	GAI Profile action MP-1.1-004 mapped to MAP-1.1	Document TEVV metrics, success criteria, and out of scope use cases for the GAI system.
GAI-MP-1.2-001	GAI Profile action MP-1.2-001 mapped to MAP-1.2	Ensure interdisciplinary GAI actor participation reflecting demographic diversity and broad domain expertise.
GAI-MP-1.2-002	GAI Profile action MP-1.2-002 mapped to MAP-1.2	Document interdisciplinary collaboration practices and their influence on GAI design choices.
GAI-MP-2.1-001	GAI Profile action MP-2.1-001 mapped to MAP-2.1	Define the specific GAI tasks and methods to be supported including generation modalities and downstream applications.
GAI-MP-2.1-002	GAI Profile action MP-2.1-002 mapped to MAP-2.1	Document training and fine-tuning approach, base model lineage, and intended deployment surface for the GAI system.
GAI-MP-2.2-001	GAI Profile action MP-2.2-001 mapped to MAP-2.2	Document GAI system knowledge cut-off, capability limits, and known failure modes.
GAI-MP-2.2-002	GAI Profile action MP-2.2-002 mapped to MAP-2.2	Document how GAI output is utilized and overseen by humans, including review thresholds and escalation criteria.
GAI-MP-2.3-001	GAI Profile action MP-2.3-001 mapped to MAP-2.3	Address scientific integrity and TEVV considerations for GAI including experimental design and validity.
GAI-MP-2.3-002	GAI Profile action MP-2.3-002 mapped to MAP-2.3	Document training data provenance, representativeness, and known biases in the data corpus.
GAI-MP-2.3-003	GAI Profile action MP-2.3-003 mapped to MAP-2.3	Document construct validation approach for GAI evaluation including benchmark relevance.
GAI-MP-2.3-004	GAI Profile action MP-2.3-004 mapped to MAP-2.3	Document evaluation cadence triggered by retraining, fine-tuning, prompt template changes, or context shifts.
GAI-MP-2.3-005	GAI Profile action MP-2.3-005 mapped to MAP-2.3	Document procedures for handling synthetic data and the risks it poses to construct validity.
GAI-MP-3.4-001	GAI Profile action MP-3.4-001 mapped to MAP-3.4	Define operator and practitioner proficiency requirements for the GAI system.
GAI-MP-3.4-002	GAI Profile action MP-3.4-002 mapped to MAP-3.4	Document operator training on prompt engineering, output review, and refusal handling.
GAI-MP-3.4-003	GAI Profile action MP-3.4-003 mapped to MAP-3.4	Document operator training on bias, fairness, and harmful content recognition.
GAI-MP-3.4-004	GAI Profile action MP-3.4-004 mapped to MAP-3.4	Document operator training on data privacy and sensitive data handling in prompts and outputs.
GAI-MP-3.4-005	GAI Profile action MP-3.4-005 mapped to MAP-3.4	Document operator training on content provenance, attribution, and disclosure requirements.
GAI-MP-3.4-006	GAI Profile action MP-3.4-006 mapped to MAP-3.4	Document operator training refresh cadence and competency assessment.
GAI-MP-4.1-001	GAI Profile action MP-4.1-001 mapped to MAP-4.1	Map legal risks of GAI components including third-party data with attention to data privacy laws.
GAI-MP-4.1-002	GAI Profile action MP-4.1-002 mapped to MAP-4.1	Map intellectual property risks including training data copyright and output similarity to protected works.
GAI-MP-4.1-003	GAI Profile action MP-4.1-003 mapped to MAP-4.1	Engage software development, IT governance, legal, compliance, and risk functions in GAI component risk mapping.
GAI-MP-4.1-004	GAI Profile action MP-4.1-004 mapped to MAP-4.1	Map risks of obscene, degrading, or abusive content from GAI components.
GAI-MP-4.1-005	GAI Profile action MP-4.1-005 mapped to MAP-4.1	Map risks of GAI components being used to produce illegal content, offensive cyber capabilities, or training data exfilt...
GAI-MP-4.1-006	GAI Profile action MP-4.1-006 mapped to MAP-4.1	Map risks of GAI hallucinated outputs damaging reputation, contracts, or regulatory compliance.
GAI-MP-4.1-007	GAI Profile action MP-4.1-007 mapped to MAP-4.1	Map risks of GAI components depending on opaque pre-trained models with unknown lineage.
GAI-MP-4.1-008	GAI Profile action MP-4.1-008 mapped to MAP-4.1	Map risks of harmful bias and homogenization including dangerous outputs targeting specific groups.
GAI-MP-4.1-009	GAI Profile action MP-4.1-009 mapped to MAP-4.1	Map data privacy risks including re-identification, memorization, and inadvertent disclosure of personal data.
GAI-MP-4.1-010	GAI Profile action MP-4.1-010 mapped to MAP-4.1	Map IP and data privacy risks of using proprietary or sensitive data in training, fine-tuning, or prompts.
GAI-MP-5.1-001	GAI Profile action MP-5.1-001 mapped to MAP-5.1	Identify likelihood and magnitude of GAI impacts both beneficial and harmful using historical data, incident reports, an...
GAI-MP-5.1-002	GAI Profile action MP-5.1-002 mapped to MAP-5.1	Document impacts on individuals, groups, communities, organizations, society, and the planet from GAI use.
GAI-MP-5.1-003	GAI Profile action MP-5.1-003 mapped to MAP-5.1	Document impacts on human-AI configurations including automation bias, deskilling, and emotional reliance.
GAI-MP-5.1-004	GAI Profile action MP-5.1-004 mapped to MAP-5.1	Document risks of dangerous, violent, or hateful content propagation and amplification through GAI.
GAI-MP-5.1-005	GAI Profile action MP-5.1-005 mapped to MAP-5.1	Document information security risks introduced by GAI deployment including prompt injection and data leakage.
GAI-MP-5.1-006	GAI Profile action MP-5.1-006 mapped to MAP-5.1	Document risks of GAI manipulating, generating, or distributing content in ways that mislead or coerce, including known...
GAI-MP-5.2-001	GAI Profile action MP-5.2-001 mapped to MAP-5.2	Engage downstream AI actors regularly to surface unanticipated GAI impacts.
GAI-MP-5.2-002	GAI Profile action MP-5.2-002 mapped to MAP-5.2	Engage with third-party data and algorithm providers to review and evaluate unanticipated impacts and shared risks.
GAI-MS-1.1-001	GAI Profile action MS-1.1-001 mapped to MEASURE-1.1	Select measurement approaches and metrics for the most significant GAI risks identified in MAP.
GAI-MS-1.1-002	GAI Profile action MS-1.1-002 mapped to MEASURE-1.1	Document GAI risks and trustworthiness characteristics that cannot be measured and the rationale.
GAI-MS-1.1-003	GAI Profile action MS-1.1-003 mapped to MEASURE-1.1	Select metrics for confabulation including factuality, citation accuracy, and hallucination rate.
GAI-MS-1.1-004	GAI Profile action MS-1.1-004 mapped to MEASURE-1.1	Select metrics for harmful bias including disaggregated performance across populations.
GAI-MS-1.1-005	GAI Profile action MS-1.1-005 mapped to MEASURE-1.1	Select metrics for information security including resistance to prompt injection and data exfiltration.
GAI-MS-1.1-006	GAI Profile action MS-1.1-006 mapped to MEASURE-1.1	Select metrics for information integrity including provenance, watermark survivability, and disclosure adherence.
GAI-MS-1.1-007	GAI Profile action MS-1.1-007 mapped to MEASURE-1.1	Select metrics for human-AI configuration including override frequency and disagreement rate.
GAI-MS-1.1-008	GAI Profile action MS-1.1-008 mapped to MEASURE-1.1	Select metrics for environmental impact including compute energy and water use across training and inference.
GAI-MS-1.1-009	GAI Profile action MS-1.1-009 mapped to MEASURE-1.1	Document metric selection rationale, expected ranges, and acceptance thresholds.
GAI-MS-1.3-001	GAI Profile action MS-1.3-001 mapped to MEASURE-1.3	Involve independent or domain experts in regular GAI assessments and updates.
GAI-MS-1.3-002	GAI Profile action MS-1.3-002 mapped to MEASURE-1.3	Document external red team or audit findings and remediation status.
GAI-MS-1.3-003	GAI Profile action MS-1.3-003 mapped to MEASURE-1.3	Consult domain experts, affected community representatives, and end users in GAI evaluations.
GAI-MS-2.10-001	GAI Profile action MS-2.10-001 mapped to MEASURE-2.10	Examine GAI privacy risk including memorization, re-identification, and leakage of personal data.
GAI-MS-2.10-002	GAI Profile action MS-2.10-002 mapped to MEASURE-2.10	Apply privacy preserving techniques such as differential privacy, data minimization, and filtered training data.
GAI-MS-2.10-003	GAI Profile action MS-2.10-003 mapped to MEASURE-2.10	Document GAI privacy assessment results, residual risk, and consent posture.
GAI-MS-2.11-001	GAI Profile action MS-2.11-001 mapped to MEASURE-2.11	Evaluate GAI fairness and harmful bias across protected and vulnerable populations.
GAI-MS-2.11-002	GAI Profile action MS-2.11-002 mapped to MEASURE-2.11	Evaluate GAI for homogenization risk where outputs converge and reduce diversity.
GAI-MS-2.11-003	GAI Profile action MS-2.11-003 mapped to MEASURE-2.11	Document GAI bias evaluation methods, datasets, and results.
GAI-MS-2.11-004	GAI Profile action MS-2.11-004 mapped to MEASURE-2.11	Plan remediation for identified GAI bias including training data balancing and post-processing controls.
GAI-MS-2.11-005	GAI Profile action MS-2.11-005 mapped to MEASURE-2.11	Re-evaluate GAI fairness after model updates, fine-tuning, or material context changes.
GAI-MS-2.12-001	GAI Profile action MS-2.12-001 mapped to MEASURE-2.12	Assess environmental impact of GAI model training including energy, water, and carbon emissions.
GAI-MS-2.12-002	GAI Profile action MS-2.12-002 mapped to MEASURE-2.12	Assess environmental impact of GAI inference at expected scale.
GAI-MS-2.12-003	GAI Profile action MS-2.12-003 mapped to MEASURE-2.12	Document sustainability practices for GAI including model efficiency and hardware reuse.
GAI-MS-2.12-004	GAI Profile action MS-2.12-004 mapped to MEASURE-2.12	Disclose GAI environmental metrics to stakeholders where required.
GAI-MS-2.13-001	GAI Profile action MS-2.13-001 mapped to MEASURE-2.13	Evaluate effectiveness of GAI TEVV metrics and processes and update them as needed.
GAI-MS-2.2-001	GAI Profile action MS-2.2-001 mapped to MEASURE-2.2	Ensure GAI human subject evaluations meet applicable requirements addressing information security and harmful bias.
GAI-MS-2.2-002	GAI Profile action MS-2.2-002 mapped to MEASURE-2.2	Apply information integrity and information security safeguards in human evaluation protocols.
GAI-MS-2.2-003	GAI Profile action MS-2.2-003 mapped to MEASURE-2.2	Apply human-AI configuration and information integrity safeguards in evaluation panel design.
GAI-MS-2.2-004	GAI Profile action MS-2.2-004 mapped to MEASURE-2.2	Use privacy enhancing technologies to minimize risks associated with linking AI-generated outputs to evaluators.
GAI-MS-2.3-001	GAI Profile action MS-2.3-001 mapped to MEASURE-2.3	Measure GAI system performance qualitatively and quantitatively under conditions similar to deployment.
GAI-MS-2.3-002	GAI Profile action MS-2.3-002 mapped to MEASURE-2.3	Test GAI under realistic load, prompt distribution, and adversarial inputs.
GAI-MS-2.3-003	GAI Profile action MS-2.3-003 mapped to MEASURE-2.3	Document GAI performance variation across user populations, languages, and domains.
GAI-MS-2.3-004	GAI Profile action MS-2.3-004 mapped to MEASURE-2.3	Test GAI behavior at the limits of its knowledge and stated capability.
GAI-MS-2.5-001	GAI Profile action MS-2.5-001 mapped to MEASURE-2.5	Demonstrate validity and reliability of the deployed GAI system through repeatable testing.
GAI-MS-2.5-002	GAI Profile action MS-2.5-002 mapped to MEASURE-2.5	Document limitations of GAI generalizability beyond development conditions.
GAI-MS-2.5-003	GAI Profile action MS-2.5-003 mapped to MEASURE-2.5	Evaluate GAI output stability across prompt phrasing, ordering, and session context.
GAI-MS-2.5-004	GAI Profile action MS-2.5-004 mapped to MEASURE-2.5	Evaluate GAI consistency over time across model versions and fine-tuning iterations.
GAI-MS-2.5-005	GAI Profile action MS-2.5-005 mapped to MEASURE-2.5	Document GAI calibration including over-confidence and refusal calibration.
GAI-MS-2.5-006	GAI Profile action MS-2.5-006 mapped to MEASURE-2.5	Document GAI evaluation results in machine readable formats with versioning.
GAI-MS-2.6-001	GAI Profile action MS-2.6-001 mapped to MEASURE-2.6	Evaluate GAI safety risks including ability to operate beyond knowledge limits and fail safely.
GAI-MS-2.6-002	GAI Profile action MS-2.6-002 mapped to MEASURE-2.6	Evaluate GAI risks of generating CBRN, dangerous, violent, or hateful content.
GAI-MS-2.6-003	GAI Profile action MS-2.6-003 mapped to MEASURE-2.6	Conduct red teaming for GAI specific harms including jailbreak, prompt injection, and content policy bypass.
GAI-MS-2.6-004	GAI Profile action MS-2.6-004 mapped to MEASURE-2.6	Implement and test safety guardrails including content filters and refusal behavior.
GAI-MS-2.6-005	GAI Profile action MS-2.6-005 mapped to MEASURE-2.6	Test GAI response to attempts to elicit prohibited content, including subtle and indirect attacks.
GAI-MS-2.6-006	GAI Profile action MS-2.6-006 mapped to MEASURE-2.6	Document residual safety risk and risk acceptance for the deployed GAI system.
GAI-MS-2.6-007	GAI Profile action MS-2.6-007 mapped to MEASURE-2.6	Test GAI behavior in safety critical contexts including healthcare, legal, and financial use.
GAI-MS-2.7-001	GAI Profile action MS-2.7-001 mapped to MEASURE-2.7	Evaluate GAI security against breaches, eavesdropping, man-in-the-middle attacks, reverse engineering, and prompt inject...
GAI-MS-2.7-002	GAI Profile action MS-2.7-002 mapped to MEASURE-2.7	Compare GAI system security posture against industry standards and best practices.
GAI-MS-2.7-003	GAI Profile action MS-2.7-003 mapped to MEASURE-2.7	Test GAI resilience to denial of service, model extraction, and data extraction attacks.
GAI-MS-2.7-004	GAI Profile action MS-2.7-004 mapped to MEASURE-2.7	Test GAI for risks of training data exfiltration and memorization disclosure.
GAI-MS-2.7-005	GAI Profile action MS-2.7-005 mapped to MEASURE-2.7	Evaluate GAI supply chain security including model registry and dependency integrity.
GAI-MS-2.7-006	GAI Profile action MS-2.7-006 mapped to MEASURE-2.7	Document GAI security testing results and remediation timelines.
GAI-MS-2.7-007	GAI Profile action MS-2.7-007 mapped to MEASURE-2.7	Evaluate GAI logging and audit trail coverage for security investigation needs.
GAI-MS-2.7-008	GAI Profile action MS-2.7-008 mapped to MEASURE-2.7	Test GAI key, credential, and secret handling for prompts and outputs.
GAI-MS-2.7-009	GAI Profile action MS-2.7-009 mapped to MEASURE-2.7	Coordinate GAI security testing with broader application and infrastructure security programs.
GAI-MS-2.8-001	GAI Profile action MS-2.8-001 mapped to MEASURE-2.8	Examine GAI transparency including disclosure that content is AI generated.
GAI-MS-2.8-002	GAI Profile action MS-2.8-002 mapped to MEASURE-2.8	Examine GAI accountability including ability to trace outputs to model, prompt, and data sources.
GAI-MS-2.8-003	GAI Profile action MS-2.8-003 mapped to MEASURE-2.8	Document GAI content provenance and watermarking approach and limitations.
GAI-MS-2.8-004	GAI Profile action MS-2.8-004 mapped to MEASURE-2.8	Document GAI disclosure obligations to users, regulators, and the public.
GAI-MS-2.9-001	GAI Profile action MS-2.9-001 mapped to MEASURE-2.9	Explain and validate the GAI model including known reasoning patterns, training data sources, and update history.
GAI-MS-2.9-002	GAI Profile action MS-2.9-002 mapped to MEASURE-2.9	Interpret GAI output within its context to inform responsible use and governance decisions.
GAI-MS-3.2-001	GAI Profile action MS-3.2-001 mapped to MEASURE-3.2	Track GAI risks where measurement is difficult including emergent capabilities and rare failure modes.
GAI-MS-3.3-001	GAI Profile action MS-3.3-001 mapped to MEASURE-3.3	Establish feedback channels for GAI end users to report problems and appeal outcomes.
GAI-MS-3.3-002	GAI Profile action MS-3.3-002 mapped to MEASURE-3.3	Establish feedback channels for affected communities not directly using the GAI system.
GAI-MS-3.3-003	GAI Profile action MS-3.3-003 mapped to MEASURE-3.3	Document GAI feedback intake, triage, and response timelines.
GAI-MS-3.3-004	GAI Profile action MS-3.3-004 mapped to MEASURE-3.3	Integrate GAI feedback insights into evaluation metrics and risk register.
GAI-MS-3.3-005	GAI Profile action MS-3.3-005 mapped to MEASURE-3.3	Publish GAI feedback outcomes and remediation status periodically.
GAI-MS-4.2-001	GAI Profile action MS-4.2-001 mapped to MEASURE-4.2	Validate GAI trustworthiness in deployment context with domain expert and end user input.
GAI-MS-4.2-002	GAI Profile action MS-4.2-002 mapped to MEASURE-4.2	Compare GAI deployment results against pre-deployment expectations and document deviations.
GAI-MS-4.2-003	GAI Profile action MS-4.2-003 mapped to MEASURE-4.2	Document GAI deployment context drift and triggers for re-evaluation.
GAI-MS-4.2-004	GAI Profile action MS-4.2-004 mapped to MEASURE-4.2	Re-validate GAI trustworthiness at defined intervals and after material change.
GAI-MS-4.2-005	GAI Profile action MS-4.2-005 mapped to MEASURE-4.2	Document GAI deployment lessons learned and route them into MAP and MEASURE refresh cycles.

Govern

19 controls

Controls in the Govern domain of NIST AI Risk Management Framework — 19 controls
Code	Title	Description
GOVERN-1.1	Legal and Regulatory Requirements	Legal and regulatory requirements involving AI are understood, managed, and documented.
GOVERN-1.2	Trustworthy AI in Policies	The characteristics of trustworthy AI are integrated into organizational policies, processes, procedures, and practices.
GOVERN-1.3	Risk Tolerance Calibration	Processes, procedures, and practices are in place to determine the needed level of risk management activities based on t...
GOVERN-1.4	Risk Management Outcomes	The risk management process and its outcomes are established through transparent policies, procedures, and other control...
GOVERN-1.5	Ongoing Monitoring and Review	Ongoing monitoring and periodic review of the risk management process and its outcomes are planned, and organizational r...
GOVERN-1.6	AI System Inventory	Mechanisms are in place to inventory AI systems and are resourced according to organizational risk priorities.
GOVERN-1.7	Decommissioning Processes	Processes and procedures are in place for decommissioning and phasing out AI systems safely and in a manner that does no...
GOVERN-2.1	Roles and Responsibilities	Roles, responsibilities, and lines of communication related to mapping, measuring, and managing AI risks are documented...
GOVERN-2.2	AI Risk Training	The organization's personnel and partners receive AI risk management training to enable them to perform their duties and...
GOVERN-2.3	Executive Accountability	Executive leadership of the organization takes responsibility for decisions about risks associated with AI system develo...
GOVERN-3.1	Diverse Decision-Making Teams	Decision-making related to mapping, measuring, and managing AI risks throughout the lifecycle is informed by a diverse t...
GOVERN-3.2	Human-AI Oversight Roles	Policies and procedures are in place to define and differentiate roles and responsibilities for human-AI configurations...
GOVERN-4.1	Critical Thinking and Safety-First Culture	Organizational policies and practices are in place to foster a critical thinking and safety-first mindset in the design,...
GOVERN-4.2	Risk and Impact Documentation	Organizational teams document the risks and potential impacts of the AI technology they design, develop, deploy, evaluat...
GOVERN-4.3	Testing and Incident Information Sharing	Organizational practices are in place to enable AI testing, identification of incidents, and information sharing.
GOVERN-5.1	External Stakeholder Feedback	Organizational policies and practices are in place to collect, consider, prioritize, and integrate feedback from those e...
GOVERN-5.2	Adjudicated Feedback Mechanisms	Mechanisms are established to enable the team that developed or deployed AI systems to regularly incorporate adjudicated...
GOVERN-6.1	Third-Party AI Risk Policies	Policies and procedures are in place that address AI risks associated with third-party entities, including risks of infr...
GOVERN-6.2	Third-Party Contingency	Contingency processes are in place to handle failures or incidents in third-party data or AI systems deemed to be high-r...

Manage

13 controls

Controls in the Manage domain of NIST AI Risk Management Framework — 13 controls
Code	Title	Description
MANAGE-1.1	Go or No-Go Determination	A determination is made as to whether the AI system achieves its intended purposes and stated objectives and whether its...
MANAGE-1.2	Risk Treatment Prioritization	Treatment of documented AI risks is prioritized based on impact, likelihood, and available resources or methods.
MANAGE-1.3	High Priority Risk Response	Responses to the AI risks deemed high priority, as identified by the MAP function, are developed, planned, and documente...
MANAGE-1.4	Residual Risk Documentation	Negative residual risks (defined as the sum of all unmitigated risks) to both downstream acquirers of AI systems and end...
MANAGE-2.1	Resources and Alternatives	Resources required to manage AI risks are taken into account, along with viable non-AI alternative systems, approaches,...
MANAGE-2.2	Sustained Value Mechanisms	Mechanisms are in place and applied to sustain the value of deployed AI systems.
MANAGE-2.3	Response to Newly Identified Risk	Procedures are followed to respond to and recover from a previously unknown risk when it is identified.
MANAGE-2.4	Supersede, Disengage, Deactivate	Mechanisms are in place and applied, and responsibilities are assigned and understood, to supersede, disengage, or deact...
MANAGE-3.1	Third-Party Monitoring	AI risks and benefits from third-party resources are regularly monitored, and risk controls are applied and documented.
MANAGE-3.2	Pre-Trained Model Monitoring	Pre-trained models which are used for development are monitored as part of AI system regular monitoring and maintenance.
MANAGE-4.1	Post-Deployment Monitoring Plans	Post-deployment AI system monitoring plans are implemented, including mechanisms for capturing and evaluating input from...
MANAGE-4.2	Continual Improvement Activities	Measurable activities for continual improvements are integrated into AI system updates and include regular engagement wi...
MANAGE-4.3	Incident Communication	Incidents and errors are communicated to relevant AI actors, including affected communities. Processes for tracking, res...

Map

18 controls

Controls in the Map domain of NIST AI Risk Management Framework — 18 controls
Code	Title	Description
MAP-1.1	Context and Intended Purpose	Intended purposes, potentially beneficial uses, context-specific laws, norms and expectations, and prospective settings...
MAP-1.2	Interdisciplinary AI Actors	Interdisciplinary AI actors, competencies, skills, and capacities for establishing context reflect demographic diversity...
MAP-1.3	Mission Alignment	The organization's mission and relevant goals for AI technology are understood and documented.
MAP-1.4	Business Value Defined	The business value or context of business use has been clearly defined or, in the case of assessing existing AI systems,...
MAP-1.5	Risk Tolerances Determined	Organizational risk tolerances are determined and documented.
MAP-1.6	System Requirements Elicitation	System requirements (e.g., the system shall respect the privacy of its users) are elicited from and understood by releva...
MAP-2.1	AI System Tasks Defined	The specific tasks and methods used to implement the tasks that the AI system will support are defined (e.g., classifier...
MAP-2.2	Knowledge Limits Documented	Information about the AI system's knowledge limits and how system output may be utilized and overseen by humans is docum...
MAP-2.3	Scientific Integrity and TEVV	Scientific integrity and TEVV considerations are identified and documented, including those related to experimental desi...
MAP-3.1	Benefits Examined	Potential benefits of intended AI system functionality and performance are examined and documented.
MAP-3.2	Costs and Errors Examined	Potential costs, including non-monetary costs, which result from expected or realized AI errors or system functionality...
MAP-3.3	Application Scope Specified	Targeted application scope is specified and documented based on the system's capability, established context, and AI sys...
MAP-3.4	Operator Proficiency	Processes for operator and practitioner proficiency with AI system performance and trustworthiness, and relevant technic...
MAP-3.5	Human Oversight Processes	Processes for human oversight are defined, assessed, and documented in accordance with organizational policies from the...
MAP-4.1	Legal Risk Mapping for Components	Approaches for mapping AI technology and legal risks of its components, including third-party data or software, are in p...
MAP-4.2	Internal Risk Controls	Internal risk controls for components of the AI system, including third-party AI technologies, are identified and docume...
MAP-5.1	Impact Likelihood and Magnitude	Likelihood and magnitude of each identified impact (both potentially beneficial and harmful) based on expected use, past...
MAP-5.2	AI Actor Engagement on Impacts	Practices and personnel for supporting regular engagement with relevant AI actors and integrating feedback about positiv...

Measure

22 controls

Controls in the Measure domain of NIST AI Risk Management Framework — 22 controls
Code	Title	Description
MEASURE-1.1	Methods and Metrics Selected	Approaches and metrics for measurement of AI risks enumerated during the MAP function are selected for implementation st...
MEASURE-1.2	Metric Appropriateness Reviews	Appropriateness of AI metrics and effectiveness of existing controls are regularly assessed and updated, including repor...
MEASURE-1.3	Independent Assessors Involved	Internal experts who did not serve as front-line developers and/or independent assessors are involved in regular assessm...
MEASURE-2.1	TEVV Documentation	Test sets, metrics, and details about the tools used during TEVV are documented.
MEASURE-2.10	Privacy Risk	Privacy risk of the AI system, as identified in the MAP function, is examined and documented.
MEASURE-2.11	Fairness and Bias	Fairness and bias, as identified in the MAP function, are evaluated and results are documented.
MEASURE-2.12	Environmental Impact	Environmental impact and sustainability of AI model training and management activities, as identified in the MAP functio...
MEASURE-2.13	TEVV Effectiveness Evaluated	Effectiveness of the employed TEVV metrics and processes in the MEASURE function are evaluated and documented.
MEASURE-2.2	Human Subjects Requirements	Evaluations involving human subjects meet applicable requirements (including human subject protection) and are represent...
MEASURE-2.3	Performance Demonstrated for Deployment Conditions	AI system performance or assurance criteria are measured qualitatively or quantitatively and demonstrated for conditions...
MEASURE-2.4	Production Monitoring	The functionality and behavior of the AI system and its components, as identified in the MAP function, are monitored whe...
MEASURE-2.5	Validity and Reliability	The AI system to be deployed is demonstrated to be valid and reliable. Limitations of the generalizability beyond the co...
MEASURE-2.6	Safety Evaluation	The AI system is evaluated regularly for safety risks; the deployed system is demonstrated to be safe, its residual nega...
MEASURE-2.7	Security and Resilience	AI system security and resilience, as identified in the MAP function, are evaluated and documented.
MEASURE-2.8	Transparency and Accountability Risks	Risks associated with transparency and accountability, as identified in the MAP function, are examined and documented.
MEASURE-2.9	Model Explainability	The AI model is explained, validated, and documented, and AI system output is interpreted within its context to inform r...
MEASURE-3.1	Risk Tracking Approaches	Approaches, personnel, and documentation are in place to regularly identify and track existing, unanticipated, and emerg...
MEASURE-3.2	Hard-to-Measure Risk Settings	Risk tracking approaches are considered for settings where AI risks are difficult to assess using currently available me...
MEASURE-3.3	End User Feedback Processes	Feedback processes for end users and impacted communities to report problems and appeal system outcomes are established...
MEASURE-4.1	Deployment Context Measurement	Measurement approaches for identifying AI risks are connected to deployment contexts and informed through consultation w...
MEASURE-4.2	Trustworthiness Validation in Deployment	Measurement results regarding AI system trustworthiness in deployment contexts and across the AI lifecycle are informed...
MEASURE-4.3	Performance Trends from AI Actors	Measurable performance improvements or declines based on consultations with relevant AI actors, including affected commu...

Frequently Asked Questions

What is NIST AI Risk Management Framework?

NIST AI Risk Management Framework is a compliance framework from United States with 5 domains and 284 controls. NIST AI 100-1 AI RMF Core + AI 600-1 Generative AI Profile. Voluntary AI risk management framework. It is used by organisations to establish and maintain compliance with industry standards and regulatory requirements.

How many controls does NIST AI Risk Management Framework have?

NIST AI Risk Management Framework has 284 controls organised across 5 domains. The largest domains are Generative AI Profile (212 controls), Measure (22 controls), Govern (19 controls). Each control defines specific requirements that organisations must implement to achieve compliance.

What frameworks does NIST AI Risk Management Framework map to?

NIST AI Risk Management Framework does not currently have cross-framework mappings in our system. Check back as we continuously expand our mapping database.

How do I get started with NIST AI Risk Management Framework compliance?

Start your NIST AI Risk Management Framework compliance journey by running a self-assessment on our platform to identify your current compliance posture. Our AI advisory can answer specific questions about NIST AI Risk Management Framework requirements, and cross-framework mapping helps you leverage existing controls from other frameworks you may already comply with. Create a free account to access all 284 controls and track your progress.

Start Your Compliance Journey

Create a free account to run self-assessments, get AI advisory, and track your compliance progress across 769 frameworks.

Get Started Free →

Free forever — no credit card required