| Publisher | National Institute of Standards and Technology (NIST) (U.S. Department of Commerce) | |
| Document ID | NIST AI 600-1 | |
| Published | July 26, 2024 | |
| Authority | Executive Order 14110 — Safe, Secure, and Trustworthy AI § 4.1(a)(i)(A) | |
| Type | Voluntary cross-sectoral profile of [[nist-ai-rmf | AI RMF 1.0]] |
| Status | Active (not withdrawn despite EO 14110 rescission) |
NIST AI 600-1, the Generative AI Profile, is a voluntary cross-sectoral profile of the NIST AI Risk Management Framework 1.0 tailored to generative AI, published July 26, 2024. It defines 12 generative-AI (GAI) risk categories and more than 200 suggested actions across the RMF's four functions. It was commissioned under Executive Order 14110 — Safe, Secure, and Trustworthy AI § 4.1(a)(i)(A) and remains active after that order's rescission.
Scope and definitions
A profile implements the AI RMF's functions, categories, and subcategories for a specific setting. NIST AI 600-1 applies them to generative foundation models and the applications built on them. Because it is cross-sectoral, it is not tailored to a single industry or deployment context.
The profile was developed through the GAI Public Working Group, which prioritized four primary considerations: governance, content provenance, pre-deployment testing, and incident disclosure.
GAI risk categories
The profile defines 12 categories of generative-AI risk:
| # | Risk |
|---|---|
| 1 | CBRN Information or Capabilities |
| 2 | Confabulation (hallucination) |
| 3 | Dangerous, Violent, or Hateful Content |
| 4 | Data Privacy |
| 5 | Environmental Impacts |
| 6 | Harmful Bias or Homogenization |
| 7 | Human-AI Configuration (anthropomorphism, reliance) |
| 8 | Information Integrity (deepfakes, disinfo) |
| 9 | Information Security (cyber uplift) |
| 10 | Intellectual Property |
| 11 | Obscene, Degrading, and/or Abusive Content |
| 12 | Value Chain and Component Integration |
The profile states that it focuses on risks with existing empirical evidence, a framing distinct from the extinction-risk emphasis of Statement on AI Risk (CAIS) and FLI — Pause Giant AI Experiments: An Open Letter.
Suggested actions
The profile lists more than 200 suggested actions. Each action is coded by AI RMF function and subcategory (GV-, MP-, MS-, MG-) and tagged with the GAI risks it addresses.
| Function | Focus |
|---|---|
| GOVERN (GV-) | Culture, policies, roles, IP and third-party governance |
| MAP (MP-) | Context, impact assessment, stakeholder engagement |
| MEASURE (MS-) | Benchmarks, red-teaming, incident tracking, provenance |
| MANAGE (MG-) | Prioritization, decommissioning, monitoring, feedback |
Status and continued use
Executive Order 14110 — Safe, Secure, and Trustworthy AI was rescinded January 20, 2025, but NIST AI 600-1 has not been withdrawn. It remains the federal voluntary baseline for generative-AI risk management and continues to be incorporated by reference in state laws and private governance programs, including California SB 53 — Transparency in Frontier AI Act, Colorado AI Act (SB 24-205) and SB 25B-004 (Date Amendment), and State AG AI Guidances (CA, NJ, MA, OR).
NIST standards are voluntary by statute; mandatory requirements such as SB 53 and the EU AI Act (Regulation 2024/1689) layer obligations on top of the profile rather than replacing it. Although the profile has not been withdrawn, the NIST AISI and GAI working-group structure could be affected by future political decisions, and Executive Order 14365 — Ensuring a National Policy Framework for AI and America's AI Action Plan signal reduced federal emphasis on pre-deployment safety standards.
NIST's AI guidance line continued after the rescission. The NIST AI Agent Standards Initiative (February 17, 2026), housed under CAISI, follows the sequence RMF 1.0 → AI 600-1 → Agent Standards. Unlike AI 600-1's 12-risk generative-AI taxonomy, the agent initiative is organized around interoperability, agent identity and authorization, and security, closer to the security-forward framing of Executive Order 14365 — Ensuring a National Policy Framework for AI and America's AI Action Plan. SP 800-53 Release 5.2.0 (August 27, 2025) adds cyber-resiliency controls (SA-24), root-cause-analysis requirements (SI-02(07)), and logging-syntax standards (SA-15(13)) relevant to agent systems operating in federal contexts.
Relation to other frameworks
| Framework | Scope | Mandatory? | Focus |
|---|---|---|---|
| NIST AI 600-1 | Generative AI | Voluntary | 12-risk taxonomy + actions |
| NIST AI Risk Management Framework (AI RMF 1.0) | General AI | Voluntary | 4-function RMF |
| EU AI Act (Regulation 2024/1689) GPAI Code | GPAI | Mandatory for systemic-risk models | Transparency, risk mgmt, cyber |
| California SB 53 — Transparency in Frontier AI Act | Large frontier developers | Mandatory | Framework publication + catastrophic-risk assessment |
| Frontier Compliance Framework (February 2026) | Frontier labs | Voluntary (industry) | Pre-deployment red-teaming |
| Anthropic's Responsible Scaling Policy (Version 3.1) / OpenAI Preparedness Framework V.2 | Single company | Voluntary (self-binding) | Capability thresholds |
Relationships
- depends-on: NIST AI Risk Management Framework (AI RMF 1.0) — this profile implements the parent framework for GAI
- related: Executive Order 14110 — Safe, Secure, and Trustworthy AI — commissioning authority (§ 4.1(a)(i)(A))
- related: NIST AI Agent Standards Initiative (Feb 2026) — post-AI-600-1 agent-era continuation of the NIST trajectory
- supports: California SB 53 — Transparency in Frontier AI Act, Colorado AI Act (SB 24-205) and SB 25B-004 (Date Amendment), State AG AI Guidances (CA, NJ, MA, OR) — state laws and guidance referencing NIST standards
- related: AI Safety Cases and Frameworks, Frontier Compliance Framework (February 2026), A Taxonomy of Systemic Risks from General-Purpose AI
- related: Anthropic's Responsible Scaling Policy (Version 3.1), OpenAI Preparedness Framework V.2
Sources
- NIST AI 600-1 source summary
- NIST AI Agent Standards Initiative (Feb 2026) — NIST AI Agent Standards Initiative (February 17, 2026)