Skip to main content

What is the NIST AI Risk Management Framework?

The AI Glossary

In our classes, we often pause to explain the ideas behind AI and the language used to describe it. These posts grew out of those conversations. They’re for anyone who would like to understand AI better, with no technical background assumed.

The NIST AI Risk Management Framework is voluntary guidance for identifying and managing risks from AI. NIST is the U.S. National Institute of Standards and Technology. Its AI RMF 1.0 was published in January 2023, with a companion profile addressing generative AI published in July 2024.

The framework is not a law, a product approval or an automatic certification. An organization can use it to structure its work, and particular contracts or policies may incorporate related requirements. That is different from claiming that the framework itself imposes a universal legal obligation.

Its core has four functions. Govern establishes responsibilities, policies and organizational processes. Map examines the context: what the system does, who is affected and where risks may arise. Measure develops and uses methods to assess those risks. Manage addresses them through decisions, controls and continuing review.

These functions are connected and iterative, not four mandatory stages completed once in sequence. New evidence can change the understanding of the use case, which may require different measurements or governance decisions. Governance supports the other functions throughout.

The framework also describes characteristics of trustworthy AI and recognizes that trade-offs depend on context. It does not prescribe one fairness metric, one monitoring frequency or one acceptable error rate for every application. Those choices need to reflect the actual consequences and available evidence.

For a clinical organization, a practical use could begin with an inventory of AI-enabled services and their owners. A team would define the referral assistant’s intended role, identify risks such as missed documents or unauthorized contacts, choose meaningful checks and agree on a response when performance deteriorates.

The important output is not a completed crosswalk claiming that every box is covered. It is a working system of decisions and evidence: who approved the use, what was tested, what remains uncertain and what would trigger a change. The framework provides a common structure for that conversation. Organizations still need to apply the laws and professional requirements relevant to their own setting.

Sources: NIST — Artificial Intelligence Risk Management Framework (AI RMF 1.0, 2023); NIST — AI RMF Core: Govern, Map, Measure and Manage; NIST — AI RMF Playbook

To learn more, take a look at the AI in Health Care Certificate of Specialization from the Advanced Learning Academy at Harvard T.H. Chan School of Public Health. You can take an individual course or work toward the full certificate.


AI is a fast-moving field. This explanation reflects our class discussions and related material at the time of writing. Terminology, capabilities and tools may change.

For suggested additions or corrections, please email me at tpanch@hsph.harvard.edu.

Dr. Trishan Panch
Program Director, AI in Health Care
Harvard T.H. Chan School of Public Health — Advanced Learning Academy

Lead with AI in Health Care

The AI Pragmatist is a monthly newsletter from ALA faculty, with new explainers & essays from the Advanced Learning Academy