OMB Memorandum M-26-04: Unbiased AI Principles
AI-generated. This page was written and translated automatically by AI and checked by software against the cited sources. No human editor reviewed it. Not legal advice.
Original title: Increasing Public Trust in Artificial Intelligence Through Unbiased AI Principles
M-26-04
Summary
This Office of Management and Budget memorandum implements Executive Order 14319 by directing agencies to require that procured large language models comply with two Unbiased AI Principles: truth-seeking and ideological neutrality. Agencies must add contractual requirements to new LLM solicitations and orders, request vendor transparency documents, and update procurement policies. The memorandum ceases to have effect after a set period unless the Director provides otherwise.
Who it applies to
Applies to each federal executive department, military department, independent establishment and wholly owned Government corporation procuring large language models. It does not apply to national security systems, though application there is encouraged. Agencies consider applying it to agency-developed LLMs and non-LLM AI models.
Key points
- New solicitations or orders for LLMs must include contractual requirements addressing compliance with the Unbiased AI Principles.
- Agencies should, where practicable, modify existing LLM contracts, at the latest before exercising any option extending performance.
- By March 11, 2026, agencies must update procurement policies and include processes for users to report outputs violating the principles.
- Agencies must request an acceptable use policy, model or system cards, end user resources and a feedback mechanism from vendors.
- The memorandum sunsets 2 years after issuance unless the Director of the Office of Management and Budget provides otherwise.
Key dates
- Signed / adopted
- Compliance date
- Official text / status page
- www.whitehouse.gov/wp-content/uploads/2025/12/M-26-04-Increasin…
www.whitehouse.gov/omb/information-resources/guidance/memoranda/
Sources are often in other languages; links open the original. - Topics
- Government use of AI · General AI governance · Transparency & labeling