ai-rmf-starter

Version: 0.1.0-alpha Maturity: Docs-first starter Basis: NIST AI RMF 1.0 (NIST AI 100-1, January 2023)

ai-rmf-starter is a practical starter kit for teams who want to operationalize the NIST AI Risk Management Framework instead of only reading it.

This repo is designed to help a team:

• define AI governance and accountability
• document intended use and misuse cases
• map benefits, harms, and affected stakeholders
• create measurable trustworthiness checks
• maintain an AI risk register and response plan
• produce current and target AI RMF profiles
• adapt sample machine-readable governance artifacts

The first worked example in this repo is based on state-policy-rag-starter, a public-sector-oriented RAG system for policy-grounded answers.

This is a strong next step after state-policy-rag-starter.

Why it fits well:

• state-policy-rag-starter shows how to build and run an AI system
• ai-rmf-starter shows how to govern, assess, and monitor that system
• together they form a better story for public sector, enterprise, and regulated environments

Related Repos

• state-policy-rag-starter Technical starter for a public-sector-oriented policy RAG system.
• state-policy-rag-ui Cross-platform UI client for the policy RAG stack.

What This Repo Covers

This starter is organized around the four NIST AI RMF core functions:

• Govern
• Map
• Measure
• Manage

It also accounts for the trustworthiness characteristics emphasized in NIST AI RMF 1.0:

• valid and reliable
• safe
• secure and resilient
• accountable and transparent
• explainable and interpretable
• privacy-enhanced
• fair, with harmful bias managed

Repo Map

• docs/FRAMEWORK_OVERVIEW.md Short explanation of the RMF and how this repo uses it.
• docs/IMPLEMENTATION_GUIDE.md A practical rollout guide for a team adopting this framework.
• docs/NIST_ALIGNMENT.md A lightweight mapping from this repo to NIST AI RMF 1.0.
• templates/system-intake.md Intake template for a new AI system.
• templates/current-profile.md Current-state AI RMF profile template.
• templates/target-profile.md Target-state AI RMF profile template.
• templates/risk-register.csv Starter risk register.
• templates/governance-checklist.md Governance and accountability checklist.
• templates/measurement-plan.md TEVV and trustworthiness measurement template.
• templates/incident-response-playbook.md Response and recovery template for AI incidents.
• examples/state-policy-rag-profile.md Stronger worked example using the state-policy-rag-starter project as the system in scope.
• examples/code/README.md Lightweight machine-readable examples for launch gates, profiles, and risk tracking.

5 Minute Quick Start

If you are new to AI governance, use this order:

1. Read docs/FRAMEWORK_OVERVIEW.md.
2. Copy templates/system-intake.md for your project.
3. Fill out templates/current-profile.md with what exists today.
4. Fill out templates/target-profile.md with what you want before production.
5. Start your risk register from templates/risk-register.csv.
6. Use templates/measurement-plan.md to define how you will test trustworthiness claims.

Good First Use Cases

This repo is especially useful for:

• RAG assistants for internal policy or knowledge retrieval
• public sector AI pilots
• procurement and architecture review packages
• internal AI review boards
• AI systems that need auditability and human oversight

Code Examples

This repo is still intentionally lightweight, but it now includes a small examples/code/ area for teams who want starter artifacts they can adapt into internal tools, CI checks, or governance workflows.

Examples include:

• a sample launch-gates YAML file
• a sample current-profile YAML file
• a sample risk-register JSON file

Important Notes

• This repo is not a certification program.
• This repo does not replace legal, privacy, security, or procurement review.
• This repo summarizes and operationalizes the NIST framework; it does not reproduce the full publication.

Source Reference

Primary source used for this starter:

• NIST AI 100-1, Artificial Intelligence Risk Management Framework (AI RMF 1.0), January 2023
• DOI: https://doi.org/10.6028/NIST.AI.100-1

Local source file used during creation:

• NIST.AI.100-1.pdf