OMB M-25-21 on federal AI use, and the “Super Intelligence” wording of Executive Order 14434, explained

M-25-21 sets how federal agencies govern their own use of AI: Chief AI Officers, governance boards, inventories and minimum practices for high-impact AI. Executive Order 14434 changes the vocabulary agencies use, not the definition behind it.

What M-25-21 is and whom it covers

OMB issued M-25-21 on April 3, 2025. It opens by citing Executive Order 14179, which the President signed on January 23, 2025, and it rescinds and replaces M-24-10, the previous memorandum on agency use of AI [1]. It is directed to the heads of all executive branch departments and agencies, including independent regulatory agencies [1]. The memo organizes its direction around three priorities, innovation, governance and public trust, and says agencies must follow the detailed implementation instructions and requirements in its appendix [1].

The memo covers an agency’s own use of AI. It defines AI by reference to section 238(g) of the John S. McCain National Defense Authorization Act for Fiscal Year 2019, and says the definition includes machine learning, reinforcement learning, transfer learning and generative AI. It excludes robotic process automation and other systems whose behavior is defined only by human-defined rules, and says no system should be considered too simple to qualify [1].

The people and bodies it requires

  • Chief AI Officer. Within 60 days of issuance, each agency head must retain or designate a CAIO. An existing official, such as a CIO, Chief Data Officer or CTO, may serve if that person has significant AI expertise. At CFO Act agencies the CAIO must be at the Senior Executive Service, Scientific and Professional or Senior Leader level, or equivalent; elsewhere at or above GS-14. Agencies must notify OMB within 30 days when the CAIO changes or the post is vacant [1].
  • CAIO duties. They include maintaining the agency’s AI use case inventory, establishing a process for determining and documenting high-impact use cases, monitoring their performance, arranging independent review before risk acceptance, tracking high-impact cases centrally, and supporting efforts to track AI spending [1].
  • AI governance board. Within 90 days, each CFO Act agency must convene a board, which may be an existing body. It is chaired at the Deputy Secretary level or equivalent with the CAIO as vice-chair, and includes offices responsible for IT, cybersecurity, data, budget, statistics, legal counsel, privacy, civil rights and civil liberties [1].
  • Chief AI Officer Council. Within 90 days, OMB convenes and chairs an interagency council whose members include the CAIOs of CFO Act agencies [1].

High-impact AI and its minimum practices

The memo defines high-impact AI as AI whose output serves as a principal basis for decisions or actions with legal, material, binding or significant effect on [1]:

  • civil rights, civil liberties or privacy;
  • access to education, housing, insurance, credit, employment and other programs;
  • access to critical government resources or services;
  • human health and safety;
  • critical infrastructure or public safety; or
  • strategic assets or resources, including high-value property and sensitive or classified information.

A high-impact determination is possible with or without human oversight. Section 6 of the memo lists categories presumed to be high-impact; an official who concludes a use case in those categories is not high-impact must document that to the CAIO [1]. For high-impact use cases, agencies must [1]:

  1. conduct pre-deployment testing and prepare risk mitigation plans, using alternative methods such as querying the service when they lack access to source code, models or data;
  2. complete an AI impact assessment before deployment covering purpose and expected benefit, data quality and fitness, impacts on privacy, civil rights and civil liberties, reassessment schedules, costs, an independent review by someone not involved in development, and a signed risk acceptance;
  3. conduct ongoing monitoring for performance and adverse impacts, including changes to the system, its context of use or its data;
  4. ensure adequate training and assessment of the people who operate the AI;
  5. provide human oversight, intervention and accountability, with a fail-safe where practicable;
  6. offer timely human review and a chance to appeal for individuals affected by AI-enabled decisions;
  7. consult and incorporate feedback from end users and the public.

Agencies have 365 days from issuance to document implementation of these practices, and must safely discontinue a high-impact AI function that does not comply. Pilots are exempt if they are limited in scale and duration, certified by the CAIO, offer opt-in and opt-out where possible, and apply the practices where practicable. A CAIO may waive a practice for a specific application after a written, risk-based determination, must recertify waivers annually, and must report waivers granted or revoked to OMB within 30 days [1]. The memo also says the risk acceptance process is separate from, and does not supersede, information system authorization under OMB Circular A-130 [1].

The rest of the deadline table

The memo states its deadlines as periods after issuance, not as calendar dates [1]:

  • 180 days: CFO Act agencies publish an AI strategy using OMB’s template; every agency submits and posts a compliance plan, or a determination that it does not use covered AI, and repeats this every two years until 2036.
  • 270 days: agencies must update internal policies on IT infrastructure, data, cybersecurity and privacy where necessary, and should develop a policy on acceptable use of generative AI.
  • Annually: agencies other than the Department of Defense and the Intelligence Community inventory their AI use cases, submit the inventory to OMB and post a public version.

OMB issued a companion memorandum on AI acquisition, M-25-22, on the same day [2].

What Executive Order 14434 changes

Executive Order 14434, “Inaugurating the Era of Super Intelligence,” was signed September 29, 2026 and published October 2, 2026 [3]. It states that, to the maximum extent permitted by law, the executive branch shall use the terms “Super Intelligence” and “SI” in place of “Artificial Intelligence” and “AI” [3]. Section 2 applies this to official correspondence, public communications, websites, reports, policy documents and other non-statutory documents [3].

Three provisions limit its reach:

  • No rewriting of old documents. Nothing in section 2 requires altering previously issued regulations, Presidential actions, contracts, grants or other historical documents [3].
  • Same technology. For the order’s purposes, “Super Intelligence” and “SI” mean the technologies and systems encompassed by “artificial intelligence” as defined in 15 U.S.C. 9401(3), unless superseded by later Presidential action or an Act of Congress [3].
  • A definition to come. Within 60 days of the order, the Assistant to the President for Science and Technology is to submit to the President proposed legislative language for a federal definition of “Super Intelligence,” including whether it should modify or supersede the statutory definition of artificial intelligence and any conforming amendments [3].

The order also says it does not impair the functions of the OMB Director relating to budgetary, administrative or legislative proposals [3]. In our reading, the order changes the words agencies use in new documents, not the obligations M-25-21 sets. One detail is worth tracking: M-25-21 defines AI by reference to the FY2019 NDAA [1], while the order points to 15 U.S.C. 9401(3) [3]. How those references are reconciled in new policies and inventories is a question for agency counsel.

Status

As checked on November 14, 2026, OMB’s memoranda page lists M-25-21, and no later entry is titled as rescinding or replacing it; the most recent entry is M-27-01, dated October 23, 2026 [2]. The deadlines above are as the memo states them; agencies and OMB may have adjusted timing through instructions we have not reviewed.

This explainer describes federal policy; it is not legal advice, and agencies and contractors should confirm with counsel how it applies to them.

Moraph’s teams help organizations with work like this. See our solutions.

Spotted an error? Tell us and we will check it and correct the explainer where needed.

  • OMB
  • M-25-21
  • Executive Order 14434
  • Chief AI Officer
  • high-impact AI
  • AI governance
  • federal agencies