Skip to content

NIST ITL — International AI Standards Landscape

Provenance & licence

Source: NIST ITL AI Program · Last observed: 2026-06-16 · Version: latest briefing — ITL AI Standards Landscape webinar (2026-03-06) · Status: planned · Licence: U.S. Government work — public domain (public-domain)

Summary

This entry tracks the NIST Information Technology Laboratory (ITL) AI Program's view of the international AI standards landscape — its role, priorities, and progress. Unlike the other pages here, it is not a single standard but a meta-source / inventory: it maps how the major standards bodies (ISO/IEC JTC 1/SC 42, IEEE, CEN-CENELEC, and others) and frameworks fit together, and where the gaps and emerging work are. NIST ITL summarised the current state in its AI Standards Landscape webinar on 6 March 2026. Watching this source is how the library notices new standards worth adding before they are widely known.

In plain language

Our explanation, not the official text

Plain-language summary in our own words. NIST material is U.S. Government work (public domain). Not legal advice.

This isn't a standard — it's NIST's map of all the AI standards and where the gaps are. We track it as an early-warning radar: when a new standard or evaluation method appears, this is usually where it surfaces first, telling us what to add to the library next.

Key terms

  • Standards landscape — the overall inventory of who is publishing what across AI standards bodies.
  • ISO/IEC SC 42 — the subcommittee that writes the AI standards (42001, 23894, etc.).
  • Meta-source — a source about other sources, used here to spot what to watch next.

In depth (in our own words)

Our explanation

Our own-words explanation. NIST material is a U.S. Government work (public domain). Not legal advice.

It's a map, not a standard. Every other page in this library is a standard, framework or regulation you can be measured against. This one is different: it is NIST's view of the whole field — which bodies are writing which AI standards, how they relate, where they overlap, and where the gaps are. You don't "comply" with a landscape; you use it to navigate.

Who is on the map. The international AI-standards ecosystem is crowded: ISO/IEC JTC 1/SC 42 (the home of ISO/IEC 42001, 23894, 42005, 42006 and more), IEEE, CEN-CENELEC (which underpins the harmonised standards the EU AI Act will lean on), plus NIST's own measurement, evaluation and testing work and its AI-agent standards activity. NIST ITL tracks how these pieces fit and where the priorities are heading — currently evaluation/testing, generative AI and agentic systems.

Why we watch it. For a "watch" library this is the early-warning radar. New standards, profiles and evaluation methods tend to surface in this landscape view before they're widely known. Tracking it is how we decide what to add next and how we make sure the standards already in the library haven't been superseded.

How to use it in an engagement. Practically, it answers two auditor questions: "Is there a recognised standard for this AI concern?" and "Are the standards my client relies on still the current ones?" Use it as the index that points you to the right standard — then audit against that standard, not against the landscape itself.

Key Sections

  • ITL's role — NIST ITL's contribution to and coordination of AI standards.
  • The landscape — the inventory of active and emerging AI standards across bodies.
  • Priorities — where NIST is focusing (evaluation, testing, agentic AI, GenAI).
  • Progress — status of in-flight standards and profiles.

Audit-Relevant Anchors

  • Standards inventory — a defensible, government-maintained map of which standard applies to which AI concern.
  • Gap signals — early warning of emerging standards an auditor should start tracking.
  • Cross-references — pointers from the landscape into NIST AI RMF, ISO/IEC 42001/23894, and agentic-AI work.

Auditor Checklist

This is a meta-source; the checks are about keeping the rest of the library current:

  • The organisation tracks an authoritative standards inventory, not an ad-hoc list.
  • The standards it relies on are still current (no superseded editions in use).
  • Emerging standards (agentic, GenAI evaluation) are on a watch list.
  • Each material AI-risk concern maps to an applicable standard.

Cross-Framework Mapping

A landscape "where it points" map (not a control crosswalk):

Landscape area Points to
AI management systems ISO/IEC 42001, ISO/IEC 23894
Risk framework NIST AI RMF, GenAI Profile (600-1)
Evaluation / testing NIST evaluation work, ISO/IEC TR 24029
Agentic / GenAI security CSA Agentic Profile, OWASP GenAI/Agentic

Recent Changes (rolling, last 5)

Date Severity What changed
2026-06-16 baseline Initial baseline: ITL AI program landscape, anchored to the 2026-03-06 AI Standards Landscape webinar.

Sources

Public web sources only — local/private provenance is kept in a private mirror.