1. Execuve Summary
Recognizing the importance of technical standards in shaping development and use of Arficial
Intelligence (AI), the President’s October 2023 Execuve Order on the Safe, Secure, and Trustworthy
Development and Use of Arficial Intelligence (EO 14110) calls for “a coordinated effort…to drive the
development and implementaon of AI-related consensus standards, cooperaon and coordinaon, and
informaon sharing” internaonally. Specifically, the EO tasks the Secretary of Commerce to “establish a
plan for global engagement on promong and developing AI standards… guided by principles set out in
the NIST AI Risk Management Framework and United States Government Naonal Standards Strategy for
Crical and Emerging Technology” (NSSCET). This plan, prepared with broad public and private sector
input, fulfills the EO’s mandate. As called for by the EO, within 180 days from publicaon of this
document, NIST will report to the President on priority U.S. government acons undertaken pursuant to
the plan.
The scope of the plan is deliberately broad, on several dimensions:
• U.S. standards stakeholders engage with many kinds of interlocutors around the globe on
standards-related work, including standards development organizaons (SDOs), industry,
academia, civil society, and foreign governments.
• In the United States, the Federal government is just one among many parcipants in a dynamic,
private sector-led standards ecosystem. Recognizing that U.S. global leadership on AI standards
hinges on engagement by stakeholders from across this ecosystem, the plan lays out objecves,
topical priories, and acons that can be taken up not just by the Federal government but by the
full array of U.S. stakeholders in AI standards.
• The plan addresses the full lifecycle of standards-related acvies. This includes the
foundaonal technical work that is typically necessary before a formal standard can be
developed, the collaborave process of developing a consensus standard, and the development
of complementary tools to help with implemenng the standard.
• The plan addresses AI-related standards of all scopes, both “horizontal” (applicable across
sectors) and “vercal” (designed for the needs of a parcular sector).
The engagement acvies idenfied in this plan are aimed at furthering four sets of outcomes:
• Scienfically sound AI standards that are accessible and amenable to adopon. Standards are
most likely to meet these criteria when they are grounded in underpinning science and technical
experience, clear, implementable, understood as neutral and unlikely to impede innovaon, and
globally accessible in a mely fashion. “Horizontal” (non-sector-specific) AI standards can further
promote adopon by serving sectoral needs and minimizing the need for per-sector adaptaon.
• AI standards that reflect the needs and inputs of diverse global stakeholders. Standards are
most likely to achieve this if they are context-sensive, performance-based, human-centered,
and responsive to societal consideraons. Though views on societal consideraons vary, exisng
internaonal instruments can provide a starng point for finding consensus, and standards
parcipants can reflect their exisng commitments in their standards work. AI standards are
1
