United States - AI Research Strategy (2021)

Overview

The U.S. National Artificial Intelligence Research and Development (R&D) Strategic Plan is an interagency roadmap produced by the National Science and Technology Council's AI bodies to guide Federally-funded AI research priorities. The plan was first published in October 2016 and formally updated in June 2019; subsequent implementation activity accelerated in 2021 after Congress enacted the National Artificial Intelligence Initiative Act of 2020 and Federal agencies established implementing offices and task forces. There is no separate, authoritative document titled the "2021 update" to the Strategic Plan; instead, 2021 is best understood as a pivotal implementation year when the National AI Initiative Office was established and the National AI Research Resource (NAIRR) Task Force was launched to operationalize plan priorities. For the authoritative plan texts see the official archives and interagency publications (example copies: NSTC, National AI R&D Strategic Plan (2016) and National AI R&D Strategic Plan: 2019 Update).

Definitions

For the purposes of Federal AI R&D coordination: "AI" refers to systems that perform tasks that would otherwise require human intelligence, including machine learning, deep learning, and symbolic approaches; "R&D" refers to fundamental and applied research and development activities supported or carried out by Federal agencies; "NAIRR" (National Artificial Intelligence Research Resource) denotes a federated research infrastructure of compute, data, and testbeds; and the "National AI Initiative Office" (NAIIO / NAIIO) is the OSTP-based coordination office established to implement the National AI Initiative Act. These terms are used consistently across NSTC/NITRD, OSTP, and NSF materials.

Governance and Institutional Framework

Governance of the strategic plan rests on an interagency architecture: the National Science and Technology Council (NSTC) and its Select Committee on Artificial Intelligence (and subordinate subcommittees) set strategy; the Networking and Information Technology Research and Development (NITRD) program's National Coordination Office supports coordination; the National AI Initiative Office (established January 12, 2021) coordinates implementation across agencies; and individual agencies (e.g., NSF, NIST, DOE, NIH, DHS, DoD) execute programs, investments, and standards work. In 2021 Congress codified many plan functions in the National Artificial Intelligence Initiative Act; OSTP and NSF launched the NAIRR Task Force (June 2021) to recommend a governance model and roadmap for a federated national research resource. Institutional arrangements established in 2021 continue to feed into plan updates and interagency reporting mechanisms.

Key Focus Areas

The strategic plan’s priority areas — maintained across 2016, 2019, and 2023 updates — include: (1) sustained long‑term fundamental AI research (theory, algorithms, hardware‑aware AI); (2) human‑AI collaboration research to improve teaming, interpretability, and human factors; (3) ethical, legal and societal implications (ELSI) research to study fairness, bias, and societal impact; (4) safety and security research for trustworthy and robust systems; (5) development and sharing of public datasets, simulation environments, and testbeds; (6) metrics, standards, and benchmarks for measurement and evaluation; (7) workforce development to broaden participation and train AI professionals; and (8) expanded public‑private partnerships to accelerate technology transfer and scale research. The 2023 update added Strategy 9 to emphasize principled international collaboration. In 2021, operational activities concentrated on establishing the NAIIO and NAIRR Task Force to advance these focus areas in practice (2023 Update, 2019 Update).

Implementation Framework

Implementation is agency‑led with interagency coordination. Key mechanisms include NSF-driven National AI Research Institutes (to fund multi‑institution collaborations), the NAIRR pilot (to demonstrate federated compute/data access), NIST’s standards and technical guidance (including development of voluntary risk‑management tools), and OSTP's NAIIO for cross‑cutting coordination. Agencies use grant programs, intramural research, public‑private partnerships, and task forces (e.g., NAIRR Task Force launched June 2021) to operationalize objectives. Funding decisions are informed by agency budgets and cross‑agency priorities established through NSTC and NITRD coordination. The plan calls for metrics for evaluating progress (benchmarks, adoption rates, workforce indicators) and for periodic reviews leading to updates (2019 and 2023 are examples of that iterative approach).

Monitoring and Evaluation

Monitoring depends on multi‑layered reporting: individual agencies report on program outputs (grants awarded, institutes created, datasets released), the NAIIO consolidates cross‑agency progress and coordinates RFIs and stakeholder engagement, and NSTC/NITRD advises on strategy alignment. The plan encourages development of quantitative benchmarks and standards to enable reproducible evaluation. Public requests for information (RFIs) and open consultations have been used to evaluate the plan and solicit community inputs (for example, RFIs issued in 2022 and 2025 on updates and action plans). Periodic plan updates (2019 and 2023) were informed by evaluation, community input, and interagency reviews.

Penalties, Liability, and Appeals

The Strategic Plan itself is non‑regulatory guidance and does not establish criminal or civil penalties. Liability, enforcement, and appeals regarding the use or deployment of AI systems remain governed by sectoral federal law, agency regulations, and applicable state laws. The plan directs research into safety, standards, and evaluation mechanisms that inform regulatory and procurement decisions. Where agencies implement requirements (e.g., in federally funded procurements or transfers), those agency actions may carry contractual or administrative remedies; such remedies are handled through the relevant agency’s legal and administrative processes.

Relationship to Other Instruments

The Strategic Plan interfaces with multiple statutes and policy instruments: the National Artificial Intelligence Initiative Act of 2020 (codifying the Initiative and establishing the NAIIO and NAIRR Task Force), federal agency strategic and budget priorities, NIST standardization work, the AI Accountability Framework guidance for agencies, and sectoral guidance (e.g., healthcare, finance). The plan complements standards, guidance, and regulations produced by agencies and informs international engagement, standards harmonization, and multi‑lateral collaboration on AI R&D policy.

International Alignment

International collaboration is explicitly addressed in the 2023 update (Strategy 9) and was a growing subtext in 2021 implementation. The U.S. strategic approach promotes international science collaboration, standards alignment, and coordinated research on safety and trustworthiness with allies and like‑minded partners, while preserving controls where necessary for national security. The NAIRR and National AI Research Institutes incorporate collaboration mechanisms for international researchers subject to governance and access rules. Agencies coordinate with global standards bodies and forums to align benchmarks, metrics, and safety practices.

Implementation Timeline

Sources and References