NIST AI 600-1 (Generative AI Profile) for defense contractors

Defense programs are buying chat assistants, retrieval systems and code-generation tools. NIST published a generative-AI profile of the AI RMF. It is useful. It is not a DoD clause. For the suggested actions that name acquisition, contractors and information security, see NIST AI 600-1 suggested actions for generative AI on a defense program.

What NIST AI 600-1 is

NIST published the Generative Artificial Intelligence Profile on 26 July 2024. The PDF, NIST AI 600-1, calls itself a cross-sectoral profile of and companion resource for the AI RMF 1.0 for Generative AI. The AI RMF remains intended for voluntary use.

Status as of 14 October 2026: the NIST publication page still dates the profile Published July 26, 2024 and still describes it as pursuant to President Biden’s Executive Order (EO) 14110. EO 14110 is the publication’s own citation; this article does not treat 600-1 as a current executive-order mandate. The NIST AI RMF Playbook is a separate AIRC resource of suggested actions for the AI RMF Core; its banner says The AI RMF 1.0 is being updated. The Playbook will be updated after the AI RMF is revised.. The AI Verify Testing Framework lists a map to this profile dated 29 May 2025; that map is the Foundation’s statement, not a NIST adoption of AI Verify.

A profile, in NIST’s words, is an implementation of the AI RMF functions, categories, and subcategories for a specific setting, application, or technology. This one is for generative AI. It defines risks that are novel to or exacerbated by the use of GAI and then provides a set of suggested actions to help organizations govern, map, measure, and manage these risks.

The four functions stay the ones in the AI RMF Core: Govern, Map, Measure and Manage. See the NIST AI RMF explainer and DoD RAI Toolkit vs NIST AI RMF.

Risks the profile actually names

Section 2 of NIST AI 600-1 lists risks unique to or exacerbated by GAI. The first items include CBRN information or capabilities, confabulation (confidently stated false content), dangerous, violent, or hateful content, data privacy, and information security, including lowered barriers for offensive cyber capabilities.

Those are the sourced reasons a defense contractor should read the profile even though it is not a military document: CBRN access, confabulation in operational text, and offensive-cyber assistance are named risks, not inferences.

The profile also says cross-sectoral profiles can be used for activities or business processes common across sectors, such as the use of large language models (LLMs), cloud-based services, or acquisition. Suggested actions include updating due diligence processes for GAI acquisition and procurement vendor assessments and acceptable-use policies that address contractors, consultants, and other third-party personnel. When systems may raise national security risks, one suggested action is to involve national security professionals in mapping, measuring, and managing those risks.

What it does not replace

• CMMC / NIST SP 800-171. 32 CFR Part 170 is about protecting FCI and CUI on contractor systems. AI 600-1 does not assign a CMMC level. See CMMC Level 2 vs Level 3 for AI systems and CMMC and AI with CUI.
• DoD Responsible AI. The RAI Toolkit is built upon the NIST AI Risk Management Framework and Toolkit, among other sources. That sentence refers to the AI RMF, not specifically to AI 600-1. We found no DoD issuance that adopts AI 600-1 as a contract requirement.
• Aviation assurance. AI 600-1 is not a means of compliance for the FAA roadmap or EASA concept paper.

A contractor reading, labeled as analysis

Analysis, not sourced fact. A workable use of AI 600-1 on a defense program is:

1. Inventory generative systems the way GV-1.6-style actions describe, including vendor and open-source tools.
2. Map the named GAI risks that actually apply (confabulation in staff work, CUI leakage, offensive-cyber assistance) without pretending the profile is a CMMC overlay.
3. Keep CUI scoping and FedRAMP questions on the CMMC page, and keep DoD ethical-principle evidence on the AIA Toolkit listing.

Questions

Is AI 600-1 mandatory for DoD contractors?

NIST describes the AI RMF as voluntary. We found no Part 170 or DFARS clause that incorporates NIST AI 600-1.

Does the profile cover classified systems?

The document is a cross-sectoral profile. It discusses national security professionals as a suggested action, not a classified-system overlay.

Where do I start if I already use the AI RMF?

Start with the Govern, Map, Measure and Manage functions you already have, then read the suggested actions in NIST AI 600-1 that attach to those functions for GAI. The ISO 42001 vs NIST AI RMF crosswalk is the management-system pairing, not a generative-AI pairing.

The directory lists tools whose public pages name the AI RMF. That tag is not a claim that the tool implements AI 600-1.