Future of Artificial Intelligence Innovation Act of 2026 (S. 3952)

The Future of Artificial Intelligence Innovation Act of 2026 is a bill in the 119th United States Congress that would place the Center for Artificial Intelligence Standards and Innovation on a statutory footing within the National Institute of Standards and Technology, establish a federal program for testing and evaluating AI systems led jointly by NIST and the Department of Energy, and support AI research, public datasets, prize competitions and research security. Its long title describes it as a bill "[t]o establish artificial intelligence standards, metrics, and evaluation tools, to support artificial intelligence research, development, and capacity building activities, to promote innovation in the artificial intelligence industry by ensuring companies of all sizes can succeed and thrive, and for other purposes" (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026); Source: congress.gov). Title I is headed "Voluntary Artificial Intelligence Standards, Metrics, Evaluation Tools, Testbeds, and International Cooperation," and the section establishing the Center provides that nothing in it gives the NIST Director enforcement authority not already in effect before enactment. The bill is a reintroduction of S. 4178 of the 118th Congress, which the Senate Commerce Committee reported in December 2024.

Status and legislative history

Sen. Todd Young (R-IN) introduced the bill on February 26, 2026 "for himself, Ms. Cantwell, Mrs. Blackburn, and Mr. Hickenlooper"; it was read twice and referred to the Committee on Commerce, Science, and Transportation (Source: congress.gov). All three cosponsors, Maria Cantwell (D-WA), Marsha Blackburn (R-TN) and John Hickenlooper (D-CO), are original cosponsors, and none has been added since introduction. A star print was ordered on March 19, 2026. As of September 17, 2026 the bill's status remained "Introduced," with no committee vote recorded, no Congressional Budget Office cost estimate, no related bills, and no Congressional Research Service summary published; the landing page had listed the summary as in progress (Source: congress.gov; congress.gov).

Text versions. Congress.gov lists one text version, "Introduced in Senate." The GovInfo package for that version is the first star print, created March 20, 2026; no earlier printing is published, so what the star print changed cannot be established from the public record. The Legislative Counsel draft the Commerce Committee minority posted on the day of introduction names only Young and Cantwell in its sponsorship line, and its operative text matches the star print apart from capitalization, one hyphen and one quotation mark (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)).

Predecessor. The Commerce Committee minority described the February 2026 measure as a reintroduction of the bipartisan Future of AI Innovation Act, and said it builds on Cantwell and Young's earlier FUTURE of AI Act, which the release credits with creating the National AI Advisory Committee (Source: commerce.senate.gov). The predecessor, S. 4178 (Future of Artificial Intelligence Innovation Act of 2024), was introduced April 18, 2024 by Cantwell with Young, Hickenlooper and Blackburn, later joined by Sens. Roger Wicker, Ben Ray Luján, Kyrsten Sinema, Mike Rounds and Chuck Schumer. Commerce ordered it reported with an amendment in the nature of a substitute on July 31, 2024 and reported it without written report on December 18, 2024, placing it on the Senate calendar as Calendar No. 725, with no further action recorded (Source: congress.gov). FedScoop reported that the 2024 version advanced out of committee by voice vote as part of a package of AI bills (Source: fedscoop.com). S. 3952 follows the structure of the 2024 reported substitute. Its main changes are the renaming of the body in new NIST Act section 22B from the "Artificial Intelligence Safety Institute" to the Center for Artificial Intelligence Standards and Innovation, an expansion of that body's enumerated functions from seven to seventeen, and the removal of the substitute's Foundation for Standards and Metrology and of its section directing a presidential prohibition on federal agency policies that promote listed equity and disparate-impact concepts for AI systems (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)).

September 2026 markup reports. A Washington Post newsletter of September 14, 2026 reported that Commerce chair Ted Cruz wanted to hold a markup of AI legislation on September 23, according to three people familiar with the committee's plans, and that the roster "would likely include bipartisan legislation from Sens. Todd Young (R-Indiana) and Cantwell, called the 'Future of AI Innovation Act,'" according to one of them (Source: washingtonpost.com). The committee's September 16, 2026 executive session considered nine other bills, and its meetings list retrieved on September 17 showed no session scheduled for September 23 (Source: commerce.senate.gov; commerce.senate.gov).

Structure

Section 2 states a sense of Congress that policies affecting artificial intelligence should maximize the potential, development and use of artificial intelligence "to benefit all private and public stakeholders." Section 100 supplies definitions for Title I, including an artificial intelligence system as "an engineered or machine-based system" that can generate "predictions, recommendations, or decisions influencing real or virtual environments" and "is designed to operate with varying levels of autonomy," and a foundation model as "an artificial intelligence model trained on broad data at scale and is adaptable to a wide range of downstream tasks" (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)).

Key provisions

Center for Artificial Intelligence Standards and Innovation (section 101)

Section 101 inserts a new section 22B into the NIST Act, after section 22A (15 U.S.C. 278h-1). Within 90 days of enactment the Director must "establish a center on artificial intelligence within the Institute," the Center for Artificial Intelligence Standards and Innovation, whose mission "is to assist the private sector and agencies in developing voluntary best practices for the robust assessment of artificial intelligence systems, which may be contributed to or inform the work on such practices in standards development organizations" (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)). The Center exists administratively as NIST CAISI (Center for AI Standards and Innovation); this section would supply the statutory basis it currently lacks.

The Center carries out its functions in coordination with NIST's laboratories. The seventeen enumerated functions are: assessing AI systems "[u]sing publicly available or voluntarily provided information" and developing guidelines and best practices to measure and improve their secure development, deployment and use; supporting AI red-teaming and coordinating on building testbeds and test environments "with allies and international partners of the United States"; publishing physical and cybersecurity tools and voluntary guidelines against vulnerabilities and attacks on systems used to create or train models, including data poisoning and those "that may be exploited by foreign adversaries"; establishing AI blue-teaming capabilities; developing tools and voluntary guidelines for detecting synthetic content, authenticating content, tracking provenance and labeling content "such as by watermarking" (AI Content Provenance); cataloging, testing and documenting AI metrics and evaluation methodologies; coordinating with counterpart international institutions on interoperability; "[p]roducing resources for Federal agencies to conduct their own evaluations of artificial intelligence systems"; convening semiannual meetings with agencies and the private sector on evaluations and on national standards in sectors including biotechnology, agriculture and health care; examining safeguards against cyber attacks on AI systems; examining safeguards "to protect against unintended use of artificial intelligence for the purpose of developing chemical, biological, radiological, nuclear, and energy-security threats or hazards" (CBRN Uplift); providing, with the Department of Homeland Security and CISA, a toolkit for AI cybersecurity incidents; developing with the Secretary of Labor a list of high-priority occupations for AI training; developing best practices on minimum data quality for scientific data in model training; examining "vulnerabilities in the supply chain of hardware, including semiconductors and microelectronics, that are critical to enabling the development and deployment of artificial intelligence"; examining federal uses of AI against fraud; and identifying "proven, scalable, and interoperable techniques and metrics to promote the development of artificial intelligence."

Section 22B also establishes a Center for Artificial Intelligence Standards and Innovation Consortium within 180 days, drawn from academia, Federal laboratories, companies "of all sizes" including developers, deployers, evaluators and users, and civil society, consulted at least quarterly and reported on annually to Congress. It directs the Director to support "voluntary, consensus-based technical standards for testing artificial intelligence system components," covering training and operating infrastructure including cloud, training and evaluation data, trained or partially trained models, and human-in-the-loop testing. Confidential content a private contributor provides is exempt from the Freedom of Information Act under 5 U.S.C. 552(b)(3), and access to it is limited to the contributor and the Center, with aggregated, deidentified information releasable at the Director's discretion. Entities "under the ownership, control, or influence of the government of a covered nation" (as defined in 10 U.S.C. 4872) may not access any of the Center's resources.

The construction clause, section 22B(g), reads in full: "Nothing in this section shall be construed to provide the Director any enforcement authority that was not in effect on the day before the date of the enactment of the Future of Artificial Intelligence Innovation Act of 2026." The clause is scoped to section 22B, and S. 4178 carried the same clause. That clause, together with the title heading and the Center's mission of "voluntary best practices," locates the Center as a standards and evaluation body rather than a regulator; no provision of section 22B requires a developer to submit a model for assessment (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)).

Testbed program and voluntary foundation-model testing (section 102)

Section 102 is the operative testing provision. Within a year, the Under Secretary of Commerce for Standards and Technology and the Secretary of Energy, "in coordination with" the Director of the National Science Foundation, must jointly establish a testbed program partnering the National Laboratories, Federal laboratories, NIST, the National Artificial Intelligence Research Resource pilot and public and private entities "to conduct tests, evaluations, and security or vulnerability risk assessments" of AI systems, "in order to develop standards and encourage development of a third-party ecosystem." The officials may use the computing resources, testbeds and expertise of those bodies "to run tests and evaluations on the capabilities and limitations of artificial intelligence systems," and "shall prioritize assessments by identifying security vulnerabilities of artificial intelligence systems, including the establishment of and utilization of existing classified testbeds, at the National Laboratories if necessary," with respect to autonomous offensive cyber capabilities, cybersecurity vulnerabilities in the AI software ecosystem, and chemical, biological, radiological, nuclear, critical infrastructure and energy-security threats. They must also develop automated and reproducible tests, assess and minimize the computing needed for testing, and organize a hackathon, and must consider tests for AI systems "trained using primarily biological sequence data" that could contribute to "the creation of a pandemic or biological weapon, including those systems used for gene synthesis" (AI Biosecurity).

Subsection (i) establishes a "Voluntary Foundation Models Test Program" giving vendors of foundation models, and of AI virtual agents and robots incorporating them, "the opportunity to voluntarily test foundation models across a range of modalities." Contributed confidential content is limited to the contributor and NIST and is exempt from FOIA, and a rule of construction provides that nothing in the section requires any person to disclose trade secrets, confidential business information or privileged information. The officials consult industry, "including private artificial intelligence laboratories, companies of all sizes, and representatives from the United States financial sector," academia and civil society; they must develop metrics and evaluate the program within three years; and the program and the voluntary test program terminate seven years after enactment (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)).

The Commerce minority release describes this as creating "testbed programs between NIST, the National Science Foundation (NSF), the Department of Energy (DOE) and the private sector to develop security risk tools and testing environments for companies to evaluate their systems for capabilities and limitations" (Source: commerce.senate.gov), and the Washington Post described "new public-private testbeds hosted by the Department of Energy, the National Science Foundation and the National Institute for Standards and Technology to evaluate AI models" (Source: washingtonpost.com). In the text there is a single program, led jointly by NIST's Under Secretary and the Secretary of Energy with NSF in a coordinating role. See Department of Energy (DOE) and NSF — National Science Foundation (AI research funder).

Materials-science application (section 103)

Section 103 is headed "National Institute of Standards and Technology and Department of Energy testbed to identify, test, and synthesize new materials," but it does not establish a second testbed. It provides that the Secretary of Commerce, acting through the Under Secretary, and the Secretary of Energy "may use the program established under section 102(b) to advance materials science and energy storage and optimization and to support advanced manufacturing" through AI, autonomous laboratories, and AI integrated with emerging technologies "such as quantum hybrid computing and robotics." The Secretaries must ensure those technologies are supported by advanced algorithms and models, uncertainty quantification, and software and workforce tools producing benchmark data, model comparison tools and best-practice guides; must enter public-private partnerships as they jointly determine appropriate; and may draw on Manufacturing USA, the Hollings Manufacturing Extension Partnership, the National Laboratories, Federal laboratories and the private sector. Section 102 is the general AI testing and security-assessment program and carries the voluntary foundation-model test program; section 103 is a permissive authority to apply that program to materials, energy storage and advanced manufacturing (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)).

Coordination, reimbursement and reporting (sections 104–105)

Section 104 requires the Secretary of Commerce to prevent duplication with Department of Energy research entities, including the National Laboratories and the Advanced Scientific Computing Research program, and with relevant industries; requires Energy Department and National Laboratory resources provided under the subtitle to be supplied on a reimbursable basis under a reimbursable agreement, subject to waiver; and provides that nothing in the subtitle lets the Commerce Department, the Center or NSF "use monetary resources of the Department of Energy or any National Laboratory." Section 105 requires a progress report within one year on implementing sections 102 and 103, including reimbursable agreements, statements of work, schedules and deliverables.

International standards coalitions (section 111)

Section 111 addresses international coalitions on innovation, development, and alignment of standards with respect to artificial intelligence. The Under Secretary and the Secretary of Energy jointly lead information exchange and coordination among agencies, and communication to the private sector and like-minded foreign governments, on international technical standards for AI. The Under Secretary must seek alliances or coalitions with like-minded governments to support private sector-led standards development, to encourage adoption of US-developed standards by international standards organizations "and to advocate for international approaches to governance of artificial intelligence that promote innovation and counter influence from foreign adversaries," to facilitate international AI research collaboration "including data sharing, expertise, and resources," to build government-to-government infrastructure, to identify best practices for AI-model cybersecurity, and to coordinate technology protection measures with the State Department, the National Security Council and NSF. Partner countries must meet "technology trust criteria," including support for the WTO Technical Barriers to Trade principles of "transparency, openness, and consensus-based decision making," and must have intellectual property protections, safety standards, risk management approaches, research security measures and export controls. The People's Republic of China is ineligible until the US Trade Representative determines it "has come into compliance with the commitments it made in connection with its accession to the World Trade Organization," after which the officials must report to Congress with a national-security justification and a human-rights plan (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)).

Regulatory impediments report (section 121)

Under Subtitle C, "Identifying Regulatory Barriers to Innovation," section 121 directs the Comptroller General of the United States (Government Accountability Office (GAO)) to report within one year "on regulatory impediments to innovation in artificial intelligence systems," covering federal statutes and regulations that directly affect AI innovation and competition among "companies of all sizes," government adoption of AI, AI's use in agency service delivery "including towards combating fraud," "Federal laws and regulations relating to infrastructure and energy that unduly burden artificial intelligence systems," and recommendations "to increase the rate of innovation in artificial intelligence systems."

Research provisions (sections 201–202)

Section 201 adds section 5103A, "Public data for artificial intelligence systems," to the National Artificial Intelligence Initiative Act of 2020. It directs the Director of the Office of Science and Technology Policy to develop a list of priorities for federal investment in creating or improving "curated, publicly available Federal Government data for training and evaluating artificial intelligence systems" and to identify where to host the datasets. The list must prioritize public-interest data, datasets from federally funded research and datasets unlikely to attract private support, and "the Director shall identify 20 datasets to be prioritized." Considerations include the priorities of America's AI Action Plan of July 2025, data representative of the entire US population, and national security threats from release; no data sharing is permitted with countries the Commerce Secretary determines are engaged in conduct detrimental to US national security or foreign policy. Datasets may be made available to the NAIRR pilot and Energy Department programs "in accordance with Executive Order 14110," an order rescinded in January 2025 (Executive Order 14110 (RESCINDED)).

Section 202 establishes federal grand challenges in artificial intelligence, inserted as a new section 5107, under which OSTP and the Interagency Committee "may establish a program to award prizes" under section 24 of the Stevenson-Wydler Technology Innovation Act, and the Secretaries of Commerce and Transportation and the NSF Director may run prize competitions and challenge-based acquisitions "[s]ubject to the availability of amounts appropriated for this purpose." The initial priorities list may include fourteen categories, among them microelectronics, "explainability and mechanistic interpretability," fundamental understanding of AI in domains including "Control" and "Steerability," "safety and controls for artificial intelligence applications," advanced manufacturing, maritime vessels, border security "including solutions relevant to the detection of fentanyl," modernization of obsolete government code, and prevention of "misuse of artificial intelligence systems for malicious purposes." Prize recipients must be US-incorporated with a US primary place of business, or US citizens or permanent residents, and all research must be conducted in the United States. The prize section sunsets after five years, and section 202 also directs GAO studies of federal prize competitions.

Research security and personnel (sections 301–303)

Section 301 addresses research security, applying existing research-security requirements to activities under the Act: those of subtitle D of title VI of the Research and Development, Competition, and Innovation Act (part of the CHIPS and Science Act, Public Law 117-167) and of section 223 of the FY2021 National Defense Authorization Act. Section 302 expands NIST's authority to hire critical technical experts under 15 U.S.C. 275 by raising the number of such personnel the Director may appoint from 15 to 30 and extending the authority, which under current law expires five years after August 9, 2022, to December 30, 2035 (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026); Source: law.cornell.edu). Section 303 provides for certifications and audits of temporary fellows: before performing work under the Act relating to AI and other critical and emerging technologies, a non-federal fellow, contractor or consultant and the agency head must certify that the fellow "will not perform any inherently governmental functions," and each agency's inspector general must audit annually how many temporary fellows the agency uses, "the entities paying any temporary fellow for their work for the agency," and their work.

Relation to other proposals

Several 2026 proposals would also give the Center a statutory role. The House Obernolte–Trahan discussion draft would authorize CAISI with a $100 million annual budget to administer a frontier-model transparency framework with licensed independent verification organizations, and pairs it with state-law preemption. The Self-Improving AI Monitoring Act discussion draft would amend NIST Act section 22A, beside the section 22B S. 3952 would add, to condition the Director's pre-deployment evaluation agreements. In the Senate, the Artificial Intelligence Risk Evaluation Act (S. 2938) would place a mandatory evaluation program for systems trained above 10^26 operations at the Department of Energy and bar deployment by non-participants (S. 2938 — Artificial Intelligence Risk Evaluation Act of 2025 (bill text)). S. 3952's testing provisions at NIST and the Energy Department are voluntary.

Reactions

The Commerce minority release quotes Cantwell: "This legislation brings together private sector and government experts to develop voluntary standards for AI, create new assessment tools, and conduct testing that will ensure the United States leads in AI-driven innovation and competitiveness for decades to come." It quotes Young: "One of my top priorities for federal AI policy is to ensure these technologies are developed in a manner that reflects American values and supports U.S. innovation." Supporting statements came from the Software & Information Industry Association, the Center for Data Innovation at ITIF, the Alliance for Secure AI, Americans for Responsible Innovation, TechNet, SeedAI, the Foundation for American Innovation and the Business Software Alliance (Source: commerce.senate.gov). The release's list of provisions includes two elements the introduced text does not contain, "transparency documentation standards" and "a multilateral research collaboration between scientific and academic institutions across the world"; both appeared in S. 4178 as introduced in 2024. MeriTalk repeated the second, and reported that the bill would "mandate federal science agencies to make at least 20 curated datasets publicly available," where the text directs OSTP to "identify 20 datasets to be prioritized" (Source: meritalk.com; Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)).

In the September 2026 Senate negotiation over frontier-AI legislation, the bill has been cited alongside Cantwell's position, as an alternative to the developer-self-testing approach reported to be in the Thune–Klobuchar draft. Cantwell has said that "meaningful legislation would require the most powerful AI models undergo testing by scientists and experts at our national laboratories" (Source: reuters.com; nextgov.com); S. 3952 would create a program at those laboratories under which such testing is voluntary. That negotiation, and the disagreement over who conducts pre-deployment testing, is described at Maria Cantwell, John Thune and AI Pre-Release Vetting. The Americans for Responsible Innovation (ARI) and Ted Cruz pages also record the measure in connection with the Commerce markup calendar.

Open questions

• Whether the bill receives a Commerce Committee markup, and whether it is folded into the broader frontier-AI framework under negotiation rather than advancing on its own.
• What the March 19, 2026 star print corrected; no earlier printing of the introduced text is public, and the committee's pre-introduction draft matches the star print in operative text.
• How the section 22B(g) construction clause would interact with any later statute conferring pre-deployment testing authority on the same Center.

Relationships

• instance-of: Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026) — the bill text
• supersedes: S. 4178 (118th Congress), the Future of Artificial Intelligence Innovation Act of 2024, reported by Senate Commerce on December 18, 2024
• related: NIST CAISI (Center for AI Standards and Innovation) — the body section 101 would establish in statute
• related: National Institute of Standards and Technology (NIST) — the host agency; section 302 raises its critical-technical-expert hiring cap
• related: Todd Young — sponsor
• related: Maria Cantwell — cosponsor; sponsor of the 2024 predecessor
• related: Marsha Blackburn — cosponsor
• related: John Hickenlooper — cosponsor
• related: Department of Energy (DOE), NSF — National Science Foundation (AI research funder) — co-lead and coordinating agency for the section 102 testbed program
• related: AI Pre-Release Vetting — the evaluation question the Center's functions bear on
• related: Office of Science and Technology Policy (OSTP) — section 201 assigns the public-data priorities list to its Director
• related: Government Accountability Office (GAO) — section 121 assigns the risk identification to the Comptroller General
• related: National AI Advisory Committee (NAIAC) — the sponsors say provisions were drafted from its recommendations
• related: Frontier Act / Great American AI Act (Obernolte–Trahan) — a separate House vehicle that would also codify the Center, and that additionally carries state-law preemption
• related: Artificial Intelligence Risk Evaluation Act (S. 2938) — the other Senate measure on federal AI evaluation pending in the 119th Congress
• related: Self-Improving AI Monitoring Act — would amend NIST Act § 22A, beside the § 22B this bill would add