Short answer: No single framework covers AI governance and compliance. Effective programs layer several frameworks, each doing a different job, and then name the person who makes the final decision. The core set:
- Law: the EU AI Act. Binding obligations for providers and deployers of high-risk AI systems, including human oversight under Articles 14 and 26.
- Risk framework: the NIST AI RMF. Voluntary good practice organized into four functions: Govern, Map, Measure, and Manage.
- Management system standard: ISO/IEC 42001. Certifiable requirements for running an AI management system.
- Governance standards: ISO 37000 and ISO/IEC 38507. Guidance for the governing body on authority and on the use of AI.
- Practitioner method: such as Checkpoint-Based Governance. Names who accepts, modifies, or rejects AI output at each checkpoint, with a record of the decision.
What kind of thing is each one: a law, a standard, a risk framework, or a method?
Each framework does a different job, and treating them as interchangeable leaves gaps in a compliance program. Basil C. Puglisi, MPA, surveyed the field in AI Governance Has No Formal Definition. Here Is One. and summed it up in two sentences: “Five institutional sources address governance. None define it.”
A law creates obligations. The EU AI Act binds providers and deployers of high-risk systems, including the human oversight duties in Articles 14 and 26.
A risk framework describes good practice. Basil’s analysis of the standard of care for AI governance explains that NIST “is not a regulator, and it issues no fines,” and that its AI Risk Management Framework rests on four functions: Govern, Map, Measure, and Manage.
A management system standard sets requirements an organization can be certified against. ISO/IEC 42001:2023 “specifies requirements for establishing, implementing, maintaining, and continually improving an Artificial Intelligence Management System (AIMS) within organizations.” The definition essay adds a caution: the standard “is widely called” the AI governance standard, yet “It is not. It is a management system standard.”
Governance standards address who holds authority. ISO 37000:2021 “gives guidance on the governance of organizations,” and ISO/IEC 38507:2022 “provides guidance for members of the governing body of an organization” on AI use.
Practitioner methods describe how decisions are made inside the work, which is where the human decision finally gets placed.
Where does each one put the human decision?
Each framework names the human in a different place, and the differences matter for compliance. The EU AI Act puts the duty in two locations. Article 14 requires high-risk systems to be designed so “they can be effectively overseen by natural persons during the period in which they are in use,” and Article 26 requires deployers to “assign human oversight to natural persons who have the necessary competence, training and authority, as well as the necessary support.” Under the Digital Omnibus (Regulation (EU) 2026/1744), these high-risk requirements apply from December 2, 2027 for stand-alone systems listed in Annex III and from August 2, 2028 for AI embedded in regulated products under Annex I.
NIST places responsibility at the top. Under GOVERN 2.3 of the AI RMF, “Executive leadership of the organization takes responsibility for decisions about risks associated with AI system development and deployment.” ISO/IEC 38507 addresses the governing body’s responsibilities for AI use, and ISO/IEC 42001 sets requirements for the management system the organization runs.
A draft standard shows how open the question remains. ISO/DIS 21520, an ISO/TC 258 draft standard on artificial intelligence in project, programme, and portfolio management, closed its Draft International Standard ballot on September 29, 2026. Basil’s public technical comment on ISO/DIS 21520 argues that its draft definition “establishes organizational structure but does not establish the decision-authority threshold that distinguishes governance from oversight, management, Responsible AI controls, or human participation.” The comment proposes a different definition: “AI Governance exists when a qualified human holds binding authority at specific checkpoints, with personal accountability for the outputs that pass through.”
In Basil’s book The Minds That Bend the Machine, Chapter 15 reviews Jessica Newman and explains why that kind of participation matters, describing how her public comments on the NIST framework helped shape a standard that organizations now adopt.
How do they fit together?
They fit together as layers, each covering what the others leave open. The standard of care analysis describes how NIST and ISO/IEC 42001 fit together: “Organizations can use the NIST framework as the risk-management engine inside an ISO/IEC 42001 management system, combining the framework’s substance with the standard’s certifiable form.” The law then sets the obligations both must help meet, and the governance standards tell the board what it owns.
The gap between the layers is structural. The definition essay explains why no single body closes it: “Standards bodies define management systems because that is what they certify. Risk frameworks define risk functions because that is what they measure.” It concludes that “No institution has the mandate, the incentive, or the cross-domain authority to publish a unified definition spanning all five lenses.”
That leaves organizations to name the decision maker themselves. Under Checkpoint-Based Governance, a named person with binding authority accepts, modifies, or rejects AI output at defined checkpoints, and a record shows the decision. The method’s status is stated plainly in Basil’s paper on human-AI collaboration: “CBG is in production use in the author’s own work, but it has not been independently validated, peer-reviewed, or deployed at enterprise scale.” It is offered as one possible path that sits beside the frameworks above rather than replacing them.
For an outside read on how a compliance plan fits together, Puglisi Consulting’s AI Strategy and Readiness Review tests the plan “against the facts it rests on and the outcomes it promises to measure.” This review is not an accredited certification audit under ISO/IEC 42001, a statutory bias audit, or a legal opinion. Certification under that standard remains a separate step.
Questions Readers Ask
Is ISO/IEC 42001 an AI governance standard?
It is an AI management system standard. ISO/IEC 42001 specifies requirements for establishing, implementing, maintaining, and continually improving an AI management system, and organizations can be certified against it. Governance in the sense of who holds decision authority is addressed more directly by standards such as ISO 37000 and ISO/IEC 38507.
Is the NIST AI RMF required by law?
No. NIST describes the AI Risk Management Framework as intended to be voluntary, and NIST is not a regulator and issues no fines. Courts, procurement rules, and contracts can still give it practical weight as evidence of reasonable practice, so many organizations adopt it even where no law requires it.
When do the EU AI Act’s high-risk oversight requirements apply?
Under the Digital Omnibus, the high-risk requirements, including Article 14 human oversight, apply from December 2, 2027 for stand-alone systems listed in Annex III and from August 2, 2028 for AI embedded in regulated products under Annex I. Organizations should confirm their own obligations with counsel.
Can NIST and ISO frameworks be used together?
Yes. One approach uses the NIST AI RMF as the risk management engine inside an ISO/IEC 42001 management system. NIST supplies the substance of identifying and managing risks, while ISO/IEC 42001 supplies a certifiable structure. Laws and governance standards then set obligations and board responsibilities around both.
Why is there no formal definition of AI governance?
Each institution defines what it certifies, measures, negotiates, or enforces. Standards bodies define management systems, risk frameworks define risk functions, intergovernmental bodies define principles, and regulators define obligations. No single body has the mandate to define AI governance across all of them, which leaves organizations to name decision makers themselves.
Sources
EU Artificial Intelligence Act. (n.d.). Article 14: Human oversight. https://artificialintelligenceact.eu/article/14/
EU Artificial Intelligence Act. (n.d.). Article 26: Obligations of deployers of high-risk AI systems. https://artificialintelligenceact.eu/article/26/
International Organization for Standardization. (2021). ISO 37000:2021, Governance of organizations, Guidance. https://www.iso.org/standard/65036.html
International Organization for Standardization. (2022). ISO/IEC 38507:2022, Information technology, Governance of IT, Governance implications of the use of artificial intelligence by organizations. https://www.iso.org/standard/56641.html
International Organization for Standardization. (2023). ISO/IEC 42001:2023, Information technology, Artificial intelligence, Management system. https://www.iso.org/standard/42001
International Organization for Standardization. (n.d.). ISO/DIS 21520, Project, programme and portfolio management, Artificial intelligence, Concepts, applications, and implications [Under development]. https://www.iso.org/standard/91551.html
LAWIT. (2026, October 12). AI Act after the omnibus: High-risk AI obligations are postponed. https://lawit.sk/en/changes/ai-omnibus-2026-1744/
National Institute of Standards and Technology. (2023). Artificial intelligence risk management framework (AI RMF 1.0) (NIST AI 100-1). https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf
Puglisi, B. C. (2026, March 14). AI governance has no formal definition. Here is one. basilpuglisi.com. https://basilpuglisi.com/ai-governance-has-no-formal-definition-here-is-one/
Puglisi, B. C. (2026, June 8). The standard of care: How NIST and ISO are turning voluntary AI governance into a liability defense. basilpuglisi.com. https://basilpuglisi.com/standard-of-care-ai-governance/
Puglisi, B. C. (2026, September 29). Human-AI collaboration needs more than a human in the loop. basilpuglisi.com. https://basilpuglisi.com/human-ai-collaboration/
Puglisi, B. C. (2026). The minds that bend the machine. https://basilpuglisi.com/minds-that-bend-the-machine/
Puglisi, B. C. (2026, September 15). Public technical comment on ISO/DIS 21520, Clause 3.2: Defining AI governance by decision authority. basilpuglisi.com. https://basilpuglisi.com/iso-21520-ai-governance-comment/
Puglisi, B. C. (2026, September 11). What is Checkpoint-Based Governance? Human authority. basilpuglisi.com. https://basilpuglisi.com/cbg/
Puglisi Consulting. (n.d.). Advisory reviews. https://puglisiconsulting.com/advisory-reviews/
Regulation (EU) 2026/1744 of the European Parliament and of the Council (Digital Omnibus on AI). (2026). Official Journal of the European Union. https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=OJ:L_202601744
I am not a lawyer, and this site and its advisory reviews do not provide legal advice. This is thought research and governance analysis based on public sources, cited materials, and human-AI review. It is intended to help executives, practitioners, insurers, and governance teams think more clearly about AI risk, liability exposure, and documentation practices. Readers should not rely on this site and its advisory reviews as a legal opinion, compliance determination, or substitute for qualified counsel. Any organization facing a legal, regulatory, contractual, or insurance question should consult its own attorney, broker, or professional adviser before acting.
#AIgenerated using the HAIA Ecosystem | CC BY-NC-SA 4.0
Free for personal, educational, and noncommercial research use with attribution. Commercial exploitation, paid productization, and enterprise commercialization require separate permission and licensing.
Suggested citation: Puglisi, B. C. (2026). Show me effective frameworks for managing AI governance and compliance. basilpuglisi.com.
Leave a Reply