Every AI Policy Ever · open in the interactive register

Published (non-binding, current)Non-binding · Voluntary code / commitments

White House Accord on Super Intelligence: Joint Commitment on Frontier Responsibilities

Record US-27

Issued by
The White House (President Donald J. Trump) with Google, Anthropic, Meta, OpenAI, xAI and Nvidia
Jurisdiction
United States
Level · region
National · North America
Published / adopted
2026-09-29
Language
English

Summary

Voluntary accord in which six leading AI companies accept that each frontier developer is responsible for its models' safety. They commit to four layers of control: internal monitoring of model capabilities and alignment (focused on cyber, bio and chemical risks and on models hacking or accessing systems), a dedicated internal verification team, an independent external auditor or evaluator, and an independent board committee overseeing remediation. Signatories will also meet regularly to develop safety standards.

Safety topics

general-purpose / frontier AIevaluation & testingaccountability & liabilitycatastrophic risk

Scope

Borderline scope Commitments by six AI companies that the White House brokered and the President co-signed; included as a government-brokered company commitment.

Status notes

Signed on 29 Sep 2026 by President Trump and by Sundar Pichai (Google), Dario Amodei (Anthropic), Mark Zuckerberg (Meta), Greg Brockman (OpenAI), Elon Musk (xAI) and Jensen Huang (Nvidia), the day the President signed Executive Order 14434 on 'Super Intelligence' terminology (US-26). The text sets no enforcement or reporting mechanism; each company is to partner with an independent external auditor or evaluator, and the accord says it may make sense over time to codify these steps into laws or regulations. On 1 Oct 2026 Reps. Jay Obernolte and Lori Trahan, authors of the FRONTIER Act (US-25), said the accord embraces principles at the heart of their bill (robust internal risk management, strong corporate governance and independent third-party evaluation) and that the bill would build on it and make it durable.

Sources

Secondary source: https://obernolte.house.gov/media/press-releases/obernolte-trahan-statement-white-house-ai-accord

Local copy

Partial (secondary source only). No official URL is recorded; only the secondary source was saved. Last attempt 2026-10-17 22:12:24 UTC

Secondary source (third-party copy, not published) archive/US-27/2026-10-11/secondary.html
Retrieved 2026-10-11 14:04:11 UTC
SHA-256 798af29ed00983636efdc44ff37253af05fd61bdce69e043cc175c0bc2a6beed

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. No copy of the accord on a White House site is stored. Recheck of 17 Oct 2026: the American Presidency Project's copy of the accord (UC Santa Barbara, a secondary source; stored again for this run, same text as the copy stored on 14 Oct 2026) gives the title and subtitle, the date of 29 September 2026, the full text and the signature block of the President and six company leaders; the summary, issuing_body and notes rest on it. The stored statement of Reps. Obernolte and Trahan of 1 Oct 2026 (obernolte.house.gov, unchanged on 17 Oct) confirms that leading AI companies signed the 'White House Accord' on the preceding Tuesday (29 Sep 2026) and names its principles, but gives only that short name. Press reports (IAPP, GovConWire, Washington Examiner, TechCrunch, CFR) say the text was shared on the President's Truth Social account after a White House roundtable and that the President called it 'morally binding'; this is not in a stored source and was moved here from notes, as was the remark that no copy was found on whitehouse.gov (fact sheets, releases, briefings and presidential actions checked by web search on 11 and 14 Oct 2026). The date of EO 14434 comes from the order's text stored for US-26.
Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-18: 7 field(s) supported.

History

  1. 2026-10-18scope_note: – → Commitments by six AI companies that the White House brokered and the President — added: AGENTS.md section 10 puts company commitments that governments brokered and published in scope with a scope_note; the accord's commitments are the six companies', co-signed by the President2026-10-17-recheck-north-america-1 · Claude Code (claude-opus-5-5)
  2. 2026-10-18notes: Announced 29 Sep 2026 (a Tuesday), after a White House roundtable and alongside → Signed on 29 Sep 2026 by President Trump and by Sundar Pichai (Google), Dario Am — removed hedges and research remarks ('Press reports say', 'Reported signatories', 'no copy on whitehouse.gov could be found ... were checked') and claims that are in no stored source (the roundtable, the Truth Social posting, the 'morally binding' remark, the counterpart-to-2023 comparison, 'auditors are chosen by the companies'); these go to verification.note. Signatories, the codification remark and the auditor partnership are quoted from the accord's text (APP copy stored in this run); the 1 Oct statement from the stored congressional statement2026-10-17-recheck-north-america-1 · Claude Code (claude-opus-5-5)
  3. 2026-10-18issuing_body: The White House (President Donald J. Trump) with Alphabet/Google, Anthropic, Met → The White House (President Donald J. Trump) with Google, Anthropic, Meta, OpenAI — aligned with the signature block of the accord: 'Google' (not 'Alphabet/Google') and 'Nvidia' (not 'NVIDIA'); 'XAI' in the block is rendered as the company's name, xAI2026-10-17-recheck-north-america-1 · Claude Code (claude-opus-5-5)

Related records

White House Voluntary AI Commitments (2023) United States · Superseded

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

Citation

The White House (President Donald J. Trump) with Google, Anthropic, Meta, OpenAI, xAI and Nvidia. White House Accord on Super Intelligence: Joint Commitment on Frontier Responsibilities. 2026-09-29. Status: Published (non-binding, current); verified 2026-10-18 (level C). Local copy archive/US-27/2026-10-11/secondary.html retrieved 2026-10-11 14:04:11 UTC, SHA-256 798af29ed00983636efdc44ff37253af05fd61bdce69e043cc175c0bc2a6beed. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record US-27.

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.