NIST Generative AI Profile (AI 600-1)
A NIST AI RMF companion focused on risks and mitigations specific to generative AI.
Published by National Institute of Standards and Technology
How to use it
This voluntary profile by Chloe Autio and colleagues applies the AI RMF to systems that generate text, images, and other content. It addresses issues such as confident factual errors, privacy exposures, harmful bias, and dependence on external components.
For an internal support assistant, test a realistic set of requests, including questions the assistant should decline or escalate. Decide what gets logged, who reviews failures, and what changes would require another release evaluation.
- Access
- Free publication from NIST