[Congressional Bills 119th Congress]
[From the U.S. Government Publishing Office]
[H.R. 9925 Introduced in House (IH)]
<DOC>
119th CONGRESS
2d Session
H. R. 9925
To provide for Federal oversight of the development and deployment of
frontier artificial intelligence in interstate and foreign commerce,
and for other purposes.
_______________________________________________________________________
IN THE HOUSE OF REPRESENTATIVES
July 23, 2026
Mr. Obernolte (for himself, Mrs. Trahan, Mrs. Houchin, Mr. Peters, Mr.
Scott Franklin of Florida, and Mr. Subramanyam) introduced the
following bill; which was referred to the Committee on Energy and
Commerce, and in addition to the Committee on Science, Space, and
Technology, for a period to be subsequently determined by the Speaker,
in each case for consideration of such provisions as fall within the
jurisdiction of the committee concerned
_______________________________________________________________________
A BILL
To provide for Federal oversight of the development and deployment of
frontier artificial intelligence in interstate and foreign commerce,
and for other purposes.
Be it enacted by the Senate and House of Representatives of the
United States of America in Congress assembled,
SECTION 1. SHORT TITLE.
This Act may be cited as the ``Frontier Risk Oversight, National
Transparency, Independent Evaluation, and Reporting Act'' or the
``FRONTIER Act''.
SEC. 2. DEFINITIONS.
In this Act:
(1) Acceptable levels of catastrophic risk mitigation.--The
term ``acceptable levels of catastrophic risk mitigation''
means risk mitigation adequate to ensure that the anticipated
benefits of a frontier model made available or used in
interstate or foreign commerce outweigh its level of
catastrophic risk, taking into consideration the probability
and magnitude of the model's anticipated benefits and
catastrophic risks.
(2) Affiliate.--The term ``affiliate'' means an entity that
satisfies any of the following requirements:
(A) Controls another entity through an
intermediary.
(B) Is through such intermediary controlled by such
other entity.
(C) Is through such intermediary under common
control with such other entity.
(3) AI; artificial intelligence; artificial intelligence
model.--The terms ``AI'', ``artificial intelligence'', and
``artificial intelligence model'' have the meaning given such
terms in section 5002 of the National Artificial Intelligence
Initiative Act of 2020 (15 U.S.C. 9401).
(4) AI-related development expenditures.--The term ``AI-
related development expenditures'' means amounts paid or
incurred in interstate or foreign commerce, without regard to
whether such amounts are expensed, capitalized, deducted, or
amortized for financial reporting or tax purposes, that are
attributable to artificial intelligence development, training,
fine-tuning, modification, security, or evaluation.
(5) Assessment.--The term ``assessment'' means a review
conducted by an IVO licensed under section 5 to assess the
adequacy of a very large frontier developer's frontier AI
framework, governance policies and practices, risk-monitoring,
and mitigation of detected risks for achieving acceptable
levels of catastrophic risk mitigation.
(6) Catastrophic risk.--The term ``catastrophic risk''
means--
(A) a foreseeable and material risk that a frontier
developer's development, storage, use, or deployment of
a frontier model will materially contribute to the
death of, or serious injury to, more than 50 people, or
more than $1,000,000,000 in damage to, or loss of,
property, arising from a single incident involving a
frontier model doing any of the following:
(i) providing, in the development or
release of a chemical, biological,
radiological, nuclear, or cyber weapon,
assistance that is not publicly available;
(ii) engaging in conduct with no meaningful
human oversight or intervention, that--
(I) is a cyberattack; or
(II) if such conduct is committed
by an individual, would constitute
murder, assault, extortion, or theft,
including such theft by false pretense;
or
(iii) evading control of such developer or
a user of such model; and
(B) does not include a foreseeable and material
risk from any of the following:
(i) information that a frontier model
outputs if the information is otherwise
publicly accessible in a substantially similar
form from a source other than a foundation
model;
(ii) lawful activity of the Federal
Government; or
(iii) harm caused by a frontier model in
combination with other software if such model
did not materially contribute to the harm.
(7) Critical safety incident.--The term ``critical safety
incident'' means any of the following:
(A) Unauthorized access to, modification of, or
exfiltration of, the model weights of a frontier model.
(B) Harm resulting from the materialization of a
catastrophic risk.
(C) Loss of control of such model.
(D) A frontier model using deceptive techniques
against its frontier developer to subvert the controls
or monitoring of such developer, outside the context of
an evaluation designed to elicit such behavior, in a
manner that demonstrates materially increased
catastrophic risk.
(8) Deploy.--The term ``deploy'' means--
(A) to make a frontier model available in
interstate or foreign commerce to a third party for
use, modification, copying, or combination with other
software; and
(B) does not include making such model available to
such third party for the primary purpose of developing
or evaluating such model.
(9) Foundation model.--The term ``foundation model'' means
an artificial intelligence model made available or used in
interstate or foreign commerce that is all of the following:
(A) Trained on a broad data set.
(B) Designed for generality of output.
(C) Adaptable to a wide range of distinctive tasks.
(10) Frontier ai framework.--The term ``frontier AI
framework'' means documented technical and organizational
protocols to manage, assess, and mitigate catastrophic risks.
(11) Frontier developer.--The term ``frontier developer''
means an entity operating in interstate or foreign commerce
that has trained, or initiated the training of, a frontier
model using a quantity of computing power that meets the
technical specifications set forth in subsection (l).
(12) Frontier model.--The term ``frontier model'' means a
foundation model that was trained using a quantity of computing
power greater than 10-26 integer or floating-point operations.
The quantity of computing power described in this paragraph
includes computing for the original training run and for any
subsequent fine-tuning, reinforcement learning, or other
substantial modification the developer applies to a preceding
foundation model.
(13) Imminent catastrophic risk.--The term ``imminent
catastrophic risk'' means a present or impending catastrophic
risk.
(14) IVO.--The term ``IVO'' means an independent
verification organization operating in interstate or foreign
commerce that is licensed under subsection (c) of section 5 and
the regulations prescribed under subsection (c) of section 3.
(15) Large frontier developer.--The term ``large frontier
developer'' means a frontier developer that, together with its
affiliates, collectively, during the preceding 36-month period,
determined as of the first day of each calendar month, both--
(A) had gross revenues in excess of $50,000,000;
and
(B) incurred not less than $1,000,000,000 in AI-
related development expenditures.
(16) Material modification.--The term ``material
modification'' means with respect to a frontier AI framework, a
significant change to such framework, including a significant
change in the assessment, mitigation, or management of
catastrophic risk carried out pursuant to such framework.
(17) Model weight.--The term ``model weight'' means a
numerical parameter that satisfies the following requirements:
(A) Is in a frontier model.
(B) Is adjusted through training.
(C) Facilitates the determination of how input data
for such model is transformed into an output of such
model.
(18) Property.--The term ``property'' includes--
(A) tangible or intangible property; and
(B) does not include the loss of value of equity.
(19) Secretary.--The term ``Secretary'' means the Secretary
of Commerce.
(20) Substantial modification.--The term ``substantial
modification'' means with respect to a frontier model, a
significant change in how such model is deployed, such as
enabling a new fine-tuning capability for, releasing a model
weight of, or adding a new feature to such model, that
necessitates for such model an assessment of catastrophic risk.
(21) Under secretary.--The term ``Under Secretary'' means
the Under Secretary of Commerce for AI Security, who shall be
appointed by the Secretary.
(22) Very large frontier developer.--The term ``very large
frontier developer'' means a frontier developer that, together
with its affiliates, collectively, during the preceding 36-
month period, determined as of the first day of each calendar
month, both--
(A) had gross revenues in excess of $5,000,000,000;
and
(B) incurred not less than $10,000,000,000 in AI-
related development expenditures.
SEC. 3. RULEMAKING.
(a) Procedure.--Each regulation prescribed under this Act shall be
prescribed in accordance with section 553 of title 5, United States
Code.
(b) Framework Requirements.--
(1) Deadline.--Not later than 180 days after the date of
enactment of this Act, the Under Secretary shall issue, and
thereafter update as appropriate, regulations establishing
minimum requirements for frontier AI frameworks.
(2) Contents.--Regulations issued under this subsection
shall, at a minimum, address each of the requirements of
section 4(a).
(3) Periodic review.--The Under Secretary shall review the
regulations issued under this subsection not less frequently
than once every year, and shall update such regulations as the
Under Secretary determines to be necessary or appropriate.
(c) IVO Licensing and Oversight.--Not later than 180 days after the
date of enactment of this Act, the Under Secretary shall prescribe
regulations necessary to carry out the licensing and oversight of IVOs
under section 5, which shall include the following:
(1) Conflict-of-interest and funding-transparency
requirements, including reporting requirements regarding the
IVOs' funding sources and revenue generation and self-audit
requirements regarding the IVOs' personnel and leadership to
ensure adequate independence from the artificial intelligence
industry.
(2) Procedures and requirements for entities to apply for
licensure as an IVO, including the following:
(A) Procedures and requirements for the applicant
to submit a plan to the Under Secretary detailing--
(i) the benchmarks, technologies, metrics,
and methodologies the applicant proposes using
to assess the adequacy of the frontier
developer's frontier AI framework, governance
policies and practices, risk-monitoring, and
mitigation of detected risks, to ensure
acceptable levels of catastrophic risk
mitigation;
(ii) the applicant's proposed procedure for
prescribing and assessing implementation of
corrective action to remedy any identified
deficiency in achieving acceptable levels of
catastrophic risk mitigation;
(iii) the applicant's proposed methodology
for ensuring that the benchmarks, technologies,
metrics, and methodologies proposed in clause
(i) keep pace with technological and
organizational changes in frontier model
development;
(iv) any subcontractors that the applicant
proposes to use to carry out its
responsibilities under this Act; and
(v) the technical qualifications of the
personnel the applicant proposes to use to
conduct assessments and of any subcontractor
the applicant plans to engage under subsection
(d) of section 5 to perform any portion of an
assessment.
(B) Findings required for the Under Secretary to
license an applicant as an IVO, including--
(i) that the applicant has demonstrated its
independence from the artificial intelligence
industry; and
(ii) that every element of the applicant's
plan is adequate to ensure that frontier models
assessed by the IVO achieve acceptable levels
of catastrophic risk mitigation.
(C) Findings that require the Under Secretary to
revoke an IVO's license, including--
(i) that the IVO's plan is materially
misleading or inaccurate;
(ii) that the IVO fails to adhere to its
plan in a manner that materially impairs its
ability to fulfill its responsibilities;
(iii) that a material change compromises
the IVO's independence from the artificial
intelligence industry;
(iv) that models verified by the IVO fail
to achieve or have failed to achieve acceptable
levels of catastrophic risk mitigation, unless
the IVO demonstrates that--
(I) the IVO identified the relevant
failure, deficiency, or material
weakness in an assessment report
submitted under subsection (f) of
section 5 or a supplemental report
submitted under subsection (i) of such
section, and recommended corrective
actions adequate to achieve acceptable
levels of catastrophic risk mitigation;
and
(II) such failure is attributable
to the very large frontier developer's
failure to implement such corrective
actions, and not to any deficiency in
the IVO's assessment, methods, or
performance of its obligations under
section 5;
(v) that evolution of technology renders
the IVO's methods obsolete for verifying
acceptable mitigation of catastrophic risk; and
(vi) that the IVO has failed to comply with
its obligations under section 5.
(D) Procedures for the formation of expert panels
with respect to one or more specific catastrophic risks
to advise the Under Secretary on the issuance and
revocation of IVO licenses.
(3) Minimum requirements for assessments, taking into
account relevant approaches, best practices, frameworks, and
methods.
(4) Provisions regarding circumstances mandating corrective
action or loss of license.
(5) Rules regarding the submission and processing of
assessment reports, including a secure method for submission of
assessment reports under subsection (f), subsection (g), and
subsection (i) of section 5 and other sensitive information,
and a process for securely processing sensitive information and
protecting trade secrets.
(6) Any other rules reasonably necessary to the
administration of the IVO oversight and licensing regime
described in section 5.
(d) Modification Criteria.--Not later than 180 days after the date
of enactment of this Act, the Under Secretary shall prescribe
regulations establishing criteria for determining whether a
modification to a frontier model constitutes a substantial modification
within the meaning of subsection (t) of section 2, and whether a
modification to a frontier AI framework constitutes a material
modification within the meaning of subsection (p) of section 2.
(e) Procedures for Reports, Redactions, and State Opt-In.--The
Under Secretary may prescribe regulations est