Every AI Policy Ever · open in the interactive register

Published (non-binding, current)Non-binding · Institutional mandate

Statement from U.S. Secretary of Commerce Howard Lutnick on Transforming the U.S. AI Safety Institute into the Pro-Innovation, Pro-Science U.S. Center for AI Standards and Innovation

Record US-15

Issued by
U.S. Department of Commerce / NIST
Jurisdiction
United States
Level · region
National · North America
Published / adopted
2025-06-03
Last updated
2026-09-30
Language
English

Summary

Reorganises the US AI Safety Institute as CAISI, industry's main government contact for AI testing. Its mandate: develop voluntary guidelines and standards, make voluntary agreements with developers, run unclassified evaluations of cyber, bio and chemical-weapons risks, assess US and adversary systems (including backdoors and foreign influence), coordinate with DoD, DOE, DHS and the Intelligence Community, and oppose burdensome foreign regulation.

Safety topics

safety instituteevaluation & testinggeneral-purpose / frontier AInational security & militarycatastrophic riskinternational cooperation

Open question

checker: Notes update (CAISSI renaming, EO 14434, evaluations) is supported by stored NIST pages. Title, date_published, date_last_updated and Commerce as co-issuer cannot be supported from stored text (level C); values are unchanged by the run and already raised as open questions. A person could pr Issue #137

The record shows its last checked values until a maintainer decides; the decision is recorded in its history.

Status notes

Announced 3 Jun 2025; the centre is part of NIST. After EO 14434 of 29 Sep 2026 (US-26), NIST presents it as the Center for Advancing Innovation and Standards for Super Intelligence (CAISSI) and its mandate refers to 'SI' systems, while its news items still use the name CAISI. Its published work includes the NIST AI 800-2 draft on automated benchmark evaluations (Jan 2026, US-21); evaluations of DeepSeek V4 Pro (1 May 2026), Z.ai's GLM-5.2 (17 Jul 2026) and GLM-5.3's cyber capabilities (17 Sep 2026); a preliminary assessment of Kimi K3's cyber capabilities with the UK AI Security Institute (23 Jul 2026); and key practices for incident readiness and response in SI training and evaluation (16 Oct 2026). America's AI Action Plan (Jul 2025, US-16) assigns it tasks that include evaluating frontier models. The House 'Great American AI Act' discussion draft (Jun 2026) would codify the centre; it has not been enacted.

Sources

Open the official source https://www.commerce.gov/news/press-releases/2025/06/statement-us-secretary-commerce-howard-lutnick-transforming-us-ai

Secondary source: https://www.nist.gov/caissi

Local copy

Partial (secondary source only). The official source could not be downloaded (HTTP 403 Forbidden); only the secondary source was saved. The browser retry also failed (Browser could not load the page: Page.goto: Timeout 45000ms exceeded.). Last attempt 2026-10-17 22:12:06 UTC

Secondary source (third-party copy, not published) archive/US-15/2026-10-17/secondary.html
Retrieved 2026-10-17 21:40:31 UTC
SHA-256 81766a303cfa04623d510f77fd35cb8193ff0b98571d96809c3c5ccee816e2e2

Verification

C
Not confirmed on an official page; the record rests on secondary sources. Check it before citing. checked 2026-10-18 · search-confirmed
CAUTION: no official source confirms the title, date and status; the record rests on secondary sources. Recheck of 17 Oct 2026: the official source, the commerce.gov statement of 3 Jun 2025, again could not be stored (HTTP 403 Forbidden; the browser retry timed out), so the title and date_published rest on the record's earlier verification of 11 Oct 2026, when the statement was read but not stored. Stored: NIST's centre page (nist.gov/caissi), on 11 Oct and again on 17 Oct 2026, and in this run NIST's super intelligence page (nist.gov/super-intelligence). The centre page confirms the CAISSI name, the mandate in the summary (now worded for 'SI' systems), the voluntary character of its agreements and standards, and the evaluations in notes (DeepSeek V4 Pro on the 11 Oct copy; the 16 Oct 2026 incident-readiness practices on the 17 Oct copy). The super intelligence page confirms that NIST does this work through CAISSI and that NIST is updating its communications under EO 14434 of 29 Sep 2026. The 30 Sep 2026 date in date_last_updated appears only in the centre page's metadata (article modified time), not in its text. The tasks in America's AI Action Plan rest on US-16's stored plan, the NIST AI 800-2 draft on US-21, and the Great American AI Act draft on US-25. Press reports (GovConWire, Sep 2026) say the centre signed pre- and post-release testing agreements with Google DeepMind, Microsoft and xAI; no stored official source confirms this.
Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-18: 3 field(s) supported.

History

  1. 2026-10-18related: [{"slug": "us-aisi-establishment", "id": "US-05"}, {"slug": "us-nist-ai-800-2", → [{"slug": "us-aisi-establishment", "id": "US-05"}, {"slug": "us-nist-ai-800-2", — two-way link added: INT-11, US-07 link(s) to US-152026-10-18-fix-related-links · Claude Code (claude-opus-5-5)
  2. 2026-10-18verification.level: A → C — level computed from the evidence by tools/pipeline.py2026-10-17-recheck-north-america-1 · Claude Code (claude-opus-5-5)
  3. 2026-10-18notes: Announced 3 Jun 2025; CAISI remains in NIST. The AI Action Plan (Jul 2025) gave → Announced 3 Jun 2025; the centre is part of NIST. After EO 14434 of 29 Sep 2026 — rewritten: the remark that the testing agreements reported by GovConWire could not be confirmed is research-process language and moved to verification.note; adds the 16 Oct 2026 incident-readiness practices now listed on the NIST page and the AI Action Plan's tasks, and names the records that other statements rest on2026-10-17-recheck-north-america-1 · Claude Code (claude-opus-5-5)

Related records

US AI Safety Institute (founding) United States · Superseded

NIST AI 800-2 United States · Draft / proposed

America's AI Action Plan United States · Published (non-binding, current)

EO 14434 (Super Intelligence terminology order) United States · In force

EO 14409 (frontier AI & cybersecurity order) United States · In force

FRONTIER Act United States · Draft / proposed

International Network of AI Safety Institutes International Network of AI Safety Institutes · Published (non-binding, current)

NIST AI 800-1 United States · Draft / proposed

Citation

U.S. Department of Commerce / NIST. Statement from U.S. Secretary of Commerce Howard Lutnick on Transforming the U.S. AI Safety Institute into the Pro-Innovation, Pro-Science U.S. Center for AI Standards and Innovation. 2025-06-03. https://www.commerce.gov/news/press-releases/2025/06/statement-us-secretary-commerce-howard-lutnick-transforming-us-ai. Status: Published (non-binding, current); verified 2026-10-18 (level C). Local copy archive/US-15/2026-10-17/secondary.html retrieved 2026-10-17 21:40:31 UTC, SHA-256 81766a303cfa04623d510f77fd35cb8193ff0b98571d96809c3c5ccee816e2e2. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record US-15.

Improve this record

Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.