The No Adversarial AI Act requires the Federal Acquisition Security Council to identify artificial intelligence systems produced or developed by foreign adversaries, publish that list publicly, and keep it current. It then directs executive agencies to review and remove listed AI from acquisition and use, relying on existing procurement-security authorities, while allowing limited exceptions for research, evaluation, counterintelligence/counterterrorism, and mission-critical needs.
This bill matters because it moves a national-security filter directly into federal AI procurement: agencies must treat designated foreign-sourced AI as presumptively ineligible, creating immediate procurement, compliance, and operational consequences for agency programs, integrators, and vendors that rely on foreign technology or supply chains tied to covered nations or entities.
At a Glance
What It Does
The bill instructs the Federal Acquisition Security Council to compile and publish a list of AI produced by foreign adversaries, update it at least every 180 days, and requires agency heads to assess and exclude listed AI from acquisition and use, using authorities in 41 U.S.C. 4713. It establishes a process to remove items from the list after owner certification and Council review.
Who It Affects
Executive agencies and their procurement officers, federal integrators and contractors supplying AI-enabled goods and services, the Federal Acquisition Security Council and OMB, and any vendor or subcontractor with ties—direct or indirect—to foreign adversary countries or persons (including 20% ownership thresholds).
Why It Matters
The statute embeds a security-first screen into federal AI sourcing and creates recurring calendar-driven listing and review obligations; it shifts risk-management from case-by-case acquisition decisions to an ongoing, centrally managed blacklist that can materially reshape program procurement options and vendor competitiveness.
More articles like this one.
A weekly email with all the latest developments on this topic.
What This Bill Actually Does
The bill assigns the Federal Acquisition Security Council the job of identifying ‘‘foreign adversary’’ AI and requires the Office of Management and Budget, in coordination with that Council, to post the resulting list online. The Council must produce an initial list quickly and then refresh it every six months; the text ties the definition of AI to existing federal statutes, and it imports the government’s prior definition of ‘‘foreign adversary’’ so designation tracks national-security listings rather than product-level safety standards.
For procurement officials, the bill makes listed products presumptively excludable: within ninety days of enactment agency heads must review current acquisitions and consider exclusion or removal of any AI supplied by an entity on the list. The statute points agencies to the authorities in 41 U.S.C. 4713—procurement-security tools such as contract exclusions, security conditions, and mitigation clauses—so agencies are expected to use existing contractual levers rather than a new enforcement regime.The bill builds in a limited reversal mechanism.
A vendor can submit a certification asserting its product is not produced or developed by a foreign adversary; the Federal Acquisition Security Council reviews that submission and must affirmatively certify removal before the product comes off the list. Separately, agency heads can approve narrow exceptions after notifying OMB and two congressional oversight committees; those exceptions are limited to scientifically valid research, testing and evaluation, counterterrorism/counterintelligence work, or to preserve mission-critical functions.Finally, the text defines ‘‘foreign adversary entity’’ broadly—covering countries designated as covered nations, entities organized or domiciled in those nations, entities with 20 percent foreign ownership by such persons, or organizations subject to direction or control by those actors—so the scope reaches many layers of corporate organization.
Implementation will therefore hinge on supplier due diligence, contract clauses, and interagency coordination between acquisition, security, and program offices.
The Five Things You Need to Know
The Federal Acquisition Security Council must develop the initial list of AI produced or developed by foreign adversaries within 60 days of enactment.
The Director of OMB must publish the list on a public website within 180 days of enactment, and the Council must update the list at least every 180 days thereafter.
Agency heads have 90 days after enactment to review and consider excluding or removing listed AI from acquisition and use and must rely on authorities in 41 U.S.C. 4713 for mitigation and exclusion.
A vendor may seek removal from the list by submitting a certification that its product is not produced or developed by a foreign adversary; the Federal Acquisition Security Council must review and certify removal before delisting.
An agency head can approve an exception to continue using a listed AI only after written notice to OMB and the Senate Homeland Security and House Oversight committees, and only for research, evaluation/testing, counterterrorism/counterintelligence, or to avoid mission-critical failure.
Section-by-Section Breakdown
Every bill we cover gets an analysis of its key sections.
Short title
Provides the Act’s short name, the 'No Adversarial AI Act.' This is a formal labeling provision and has no operational effect beyond identifying the statute for citation.
Council list creation, publication, and updates
Requires the Federal Acquisition Security Council to develop a list of any artificial intelligence produced or developed by a foreign adversary within 60 days of enactment, and directs OMB (in coordination with the Council) to publish that list within 180 days. The Council must refresh the list no less frequently than every 180 days. The section also provides a mechanism for removal from the list that starts with a vendor certification and ends with Council certification—placing the burden on the owner to demonstrate non-affiliation and on the Council to affirmatively approve delisting.
Agency review and procurement-security authorities
Directs heads of executive agencies to review and consider exclusion and removal of listed AI within 90 days, and explicitly points agencies to section 4713 of title 41 as the authorities to use for mitigation. Practically, that means agencies should apply existing procurement-security tools—such as restrictions on award, contract clauses, or other administrative remedies—rather than relying on new criminal or civil penalties in this statute.
Exceptions and mandatory notice
Creates a narrow exceptions process: an agency head may authorize continued acquisition or use of listed AI only after written notice to OMB and the two named congressional committees, and only for one of four defined reasons (scientifically valid research, evaluation/testing/analysis, counterterrorism/counterintelligence, or to avoid jeopardizing mission-critical functions). This makes exceptions available but subject to executive and congressional visibility.
Scope-defining definitions and ownership thresholds
Imports the federal definition of artificial intelligence from the National AI Initiative Act and uses the 'covered nation' concept from 10 U.S.C. 4872(f)(2) to define 'foreign adversary.' It defines 'foreign adversary entity' expansively to include entities with 20 percent foreign ownership or those under foreign direction or control, which broadens the range of organizations that could appear on the list and triggers due-diligence obligations for prime contractors and subcontractors.
This bill is one of many.
Codify tracks hundreds of bills on Defense across all five countries.
Explore Defense in Codify Search →Who Benefits and Who Bears the Cost
Every bill creates winners and losers. Here's who stands to gain and who bears the cost.
Who Benefits
- Federal national-security and counterintelligence offices — They gain a centralized mechanism to block or limit procurement of AI from entities tied to hostile states, reducing a class of supply-chain risks without requiring a series of ad hoc contracting fights.
- Domestic AI vendors and non-covered foreign suppliers — They stand to gain competitively as agencies shift procurement away from listed suppliers and toward alternatives that are not tied to covered nations or entities.
- Federal Acquisition Security Council and OMB — These offices receive formal authority and a recurring role to set and maintain a security-focused procurement boundary, increasing their influence over agency sourcing decisions.
Who Bears the Cost
- Executive agencies and program offices — They must conduct rapid reviews, apply 41 U.S.C. 4713 authorities, manage exceptions, and potentially replace or re-engineer systems that rely on listed AI, creating programmatic and budgetary strain.
- Prime contractors and integrators using third-party AI components — They will face higher compliance burdens to trace ownership, defend supply-chain assertions, and potentially redesign solutions if key components are listed.
- Vendors identified as tied to foreign adversary entities — They risk exclusion from federal business absent a successful delisting certification, and could face reputational and commercial harm even while the listing process plays out.
Key Issues
The Core Tension
The bill confronts a classic trade-off: prioritizing cybersecurity and national-security risk reduction by excluding AI tied to foreign adversaries versus preserving agency mission capability and procurement flexibility. Strengthening the security posture through a central blacklist reduces certain supply-chain risks but can also force agencies to sacrifice ready-made capabilities, impose heavy compliance costs, and potentially concentrate sourcing among fewer vendors—creating another set of operational vulnerabilities.
The bill centralizes a blacklist approach but leaves several implementation decisions unresolved. First, the statute pulls definitions from multiple prior laws (the National AI Initiative Act for 'AI' and 10 U.S.C. for 'covered nation'), which could create scope and interpretive gaps—decisions about which AI techniques or composite systems qualify may become contested and drive lengthy technical review work.
Second, the owner-driven delisting process places the onus on vendors to prove a negative (not produced or developed by a foreign adversary) and requires Council certification; that creates legal and evidentiary questions about what documentation suffices, how to adjudicate conflicting evidence, and how long delisting reviews may take.
Operationally, the exception categories are broad enough to be material (for example, 'mission-critical functions') but provide no procedural limits or oversight beyond notice to OMB and two committees; agencies could rely on exceptions frequently if alternatives are unavailable, undermining the statute’s protective intent. Finally, the statute relies on existing procurement-security authorities rather than establishing penalties or a dedicated enforcement mechanism, which speeds implementation but may leave uneven application across agencies and create winners and losers based on varying acquisition office capacity and risk tolerance.
Try it yourself.
Ask a question in plain English, or pick a topic below. Results in seconds.