NIST AI Risk Management Framework and the Generative AI Profile
Voluntary. The framework and the Generative AI Profile are NIST guidance, not legislation. They do not bind an Australian law firm unless a contract mandates them.
A voluntary framework, AI RMF 1.0, for managing AI risk, plus the July 2024 Generative AI Profile. NIST says version 1.0 is being revised.
MaintainedLast reviewed
Key facts
- Issued by
- National Institute of Standards and Technology
- Sources listed
- 3
This is a reading aid for an Australian law firm. It is not legal advice, not a certification, and not a finding that a tool or a firm complies. The sentences paraphrase the sources listed, checked on the date shown. Where a standard is sold by ISO or the AICPA, the page uses the publisher’s public description and does not reproduce the standard.
What it is
NIST released the AI Risk Management Framework on 26 January 2023. NIST says it is for voluntary use, and that it is meant to improve the ability to incorporate trustworthiness into the design, development, use and evaluation of AI products, services and systems. The functions the Generative AI Profile organises its actions under are Govern, Map, Measure and Manage.
On 26 July 2024 NIST released NIST AI 600-1, the Generative Artificial Intelligence Profile. NIST says the profile helps organisations identify risks that are unique to or exacerbated by generative AI, and proposes actions. The profile says it is a cross-sector companion to AI RMF 1.0, and that a profile is an implementation of the framework’s functions for a setting or a technology, here generative AI. It is written for voluntary use.
NIST’s framework page says AI RMF 1.0 is being revised as part of the White House AI Action Plan. This page describes the published 1.0 framework and the published July 2024 profile. It does not describe a revision that NIST has not yet issued.
Who it applies to
Both documents are voluntary and are written for organisations that design, develop, use or evaluate AI. An Australian law firm can use them as a reading list. They do not bind the firm. They do not displace the conduct rules or the Privacy Act.
What it asks of a law firm using AI
- The profile’s Govern actions include understanding legal and regulatory requirements, documenting them, and integrating trustworthiness into policies. They include policies for the origin of training and generated data, evaluation of risky capabilities, an inventory of generative systems, and a way to deactivate a system.
- Map is about understanding purpose, users, beneficial uses, legal context, and the risks of components, including third-party data and software and intellectual-property risk.
- Measure is about choosing how significant risks will be measured, and about evaluating safety, security and resilience. The profile says risks that will not or cannot be measured should be documented.
- Manage includes mechanisms to supersede, disengage or deactivate a system that performs outside its intended use.
- The profile is a set of suggested actions mapped to the framework. It is not a conformity assessment, and NIST does not certify an organisation against it.
How it maps to this framework
The rows are this site’s questions. Where the cell says “Not set here”, the instrument does not set that requirement. The gap is still recorded, so it is not treated as answered.
| This site’s question | What the instrument says that bears on it |
|---|---|
| Efficiency Whether the tool saves total human work at the quality the task needs. | Not set here. |
| Effectiveness Whether the output is good enough for the task. The catalogue does not score this. | Measure is the profile’s place for evaluating whether a system stays inside its risk tolerance, including safety and security. It does not test whether a citation is real. That check remains the lawyer’s. |
| Due diligence What the firm has actually read about hosting, training and assurance before it relies on the tool. | Map and Govern are due diligence in NIST’s language: purpose, legal context, third-party components, and an inventory. Use them as questions. Do not treat a vendor’s “aligned to the NIST AI RMF” line as an answer. |
| Confidentiality Whether client and other confidential information can go into the tool. | The profile discusses data privacy, intellectual property, and special handling of personal, privileged or sensitive data in inventory entries. Those are suggested inventory fields. They are not a permission to upload privileged material. |
| Disclosure When the firm tells the client, and where required the court or another party, that AI was used. | One suggested Govern action is to establish roles and procedures for communicating incidents and for incident disclosures, and to update those processes. That is an internal and stakeholder process the profile suggests. It is not an Australian client-disclosure rule. |
| Certifications ISO, SOC 2 and IRAP names recorded against the tool. A listed name is not a finding that the certificate covers the firm’s plan. | The AI RMF is not a certification. ISO/IEC 42001 is a different document. Do not record one as the other. |
| Data residency Where the tool holds data, including whether that place is recorded as Australia. | Not set here. |
Sources
- NIST, AI Risk Management Framework
- NIST AI 600-1, Generative Artificial Intelligence Profile (July 2024)
- NIST AI 600-1, Generative AI Profile (PDF)
Checked 2026-10-10. Issued by National Institute of Standards and Technology.