Increasing Public Trust in Artificial Intelligence Through Unbiased AI Principles (OMB Memorandum M-26-04)

Summary

OMB memorandum implementing EO 14319. Every agency solicitation or order for a large language model issued after it must include contract requirements on compliance with the Unbiased AI Principles of truth-seeking and ideological neutrality, and agencies must request at least the vendor's acceptable use policy, model, system or data cards, end-user resources and a feedback mechanism. Agencies may ask for more, such as bias evaluation results and pre- and post-training details, and their procurement policies must let users report outputs that violate the principles. National security systems are excluded.

Safety topics

Status notes

Issued 11 December 2025 under section 4 of EO 14319 (US-17); it complements OMB M-25-22 (US-13). Agencies had to update their procurement policies by 11 March 2026 and should modify existing LLM contracts, where practicable, before exercising any option that extends them. Agencies need not apply it to LLMs acquired under a free, open-source licence but should check such models against the principles and against M-25-21 (US-12) before use. Agencies should make the transparency requirements material to eligibility and payment, to support termination for default. The memorandum ceases to have effect two years after its issuance (11 December 2027) unless the OMB Director provides otherwise.

Sources

Open the official source https://www.whitehouse.gov/wp-content/uploads/2025/12/M-26-04-Increasing-Public-Trust-in-Artificial-Intelligence-Through-Unbiased-AI-Principles-1.pdf

Local copy

Verification

History

1. 2026-10-18notes: Issued 11 Dec 2025, the same day as EO 14365. Agencies had to update procurement → Issued 11 December 2025 under section 4 of EO 14319 (US-17); it complements OMB — rewritten from the memo: legal basis, deadlines, open-source due diligence, materiality and the sunset; the same-day remark about EO 14365, which the memo does not mention, removed2026-10-17-recheck-north-america-1 · Claude Code (claude-opus-5-5)
2. 2026-10-18summary: Implements EO 14319. New federal LLM solicitations must require compliance with → OMB memorandum implementing EO 14319. Every agency solicitation or order for a l — the list of minimum transparency items was incomplete (end-user resources missing); the user reporting process and the national security exclusion added2026-10-17-recheck-north-america-1 · Claude Code (claude-opus-5-5)
3. 2026-10-16related: [{"slug": "us-eo-14319", "id": "US-17"}, {"slug": "us-omb-m-25-21", "id": "US-12 → [{"slug": "us-eo-14319", "id": "US-17"}, {"slug": "us-omb-m-25-21", "id": "US-12 — two-way link to US-29 added in run 2026-10-16-discover-leads2026-10-16-discover-leads · Claude Code (claude-opus-5-5)

Related records

EO 14319 (Unbiased AI Principles) United States · In force

OMB M-25-21 United States · In force

OMB M-25-22 United States · In force

EO 13960 United States · In force

Citation

Office of Management and Budget, Executive Office of the President. Increasing Public Trust in Artificial Intelligence Through Unbiased AI Principles (OMB Memorandum M-26-04) [OMB Memorandum M-26-04]. 2025-12-11. https://www.whitehouse.gov/wp-content/uploads/2025/12/M-26-04-Increasing-Public-Trust-in-Artificial-Intelligence-Through-Unbiased-AI-Principles-1.pdf. Status: In force; verified 2026-10-18 (level A). Local copy archive/US-19/2026-10-11/official.pdf retrieved 2026-10-11 14:03:57 UTC, SHA-256 d33b0277d720baac843babc5f24269e4c0b8777447870a36309d8b4f3d9c79bc. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record US-19.

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.