Rulemaking authority related to the use of algorithmic decision-making systems

(a)

Authority

(1)

In general

The head of each agency that is authorized to enforce a Federal law is authorized to make rules (in accordance with section 553 of title 5, United States Code, notwithstanding any rulemaking procedures that are otherwise applicable to the agency) to regulate the use of algorithmic decision-making systems— (A)if the head of the agency determines, based on evidence that is available to the agency, that the use of such a system is likely to materially contribute to violations of the Federal law; and (B)for the purpose of mitigating such violations. (2)

Enforcement

A violation of a rule made under this subsection shall be treated as a violation of the Federal law described in paragraph (1) for purposes of administrative and civil enforcement of such law by the agency. (3)

Advanced notice of proposed rulemaking

(A)

In general

Except as provided in subparagraph (B), not later than 60 days prior to issuing a notice of proposed rulemaking with respect to a rulemaking under this subsection, the head of an agency shall publish an advanced notice of proposed rulemaking to solicit public comment on whether the use of the algorithmic decision-making system that the agency proposes to regulate is likely to materially contribute to violations of the Federal law that the agency is authorized to enforce. An advanced notice of proposed rulemaking issued under this subparagraph shall, as appropriate, describe— (i)the algorithmic decision-making system or category of systems at issue; (ii)the provisions of Federal law implicated; and (iii)the types of information, evidence, and public comment sought by the agency. (B)

Exception

The requirement under subparagraph (A) shall not apply if the head of the agency determines that— (i)the record of a rulemaking initiated prior to the date of enactment of this Act; or (ii) a prior advanced notice of proposed rulemaking issued pursuant to this Act,

is sufficient to support a notice of proposed rulemaking under section 553 of title 5, United States Code, with respect to the rulemaking under this section.

(b)

Consultation and coordination during rulemaking

(1)

Preliminary consultation

(A)

In general

Prior to issuing a notice of proposed rulemaking with respect to a rulemaking under this subsection, the head of an agency shall, as appropriate, consult with the Administrator of the Office of Information and Regulatory Affairs on any reasonably foreseeable overlap, inconsistency, or conflict of the proposed rule with any other rule or proposed rule of another agency to regulate an algorithmic decision-making system under this section. (B)

Summary

If the head of the agency proceeds with a rulemaking under this section, the head of the agency shall include a summary of the consultation under this paragraph in the notice of proposed rulemaking. (2)

Technical consultation

The head of an agency that is making a rule under subsection (a) shall— (A)consult, as appropriate, with the Office of Science and Technology Policy and the National Institute of Standards and Technology on the technical characteristics, evaluation, and governance of an algorithmic decision-making system that the agency seeks to regulate in its rule; and (B)include in the rulemaking record a summary of any consultation conducted under this paragraph. (3)

Coordination

The head of each agency shall, to the extent practicable, coordinate with the heads of other agencies to ensure that rules made under subsection (a) do not impose conflicting requirements regarding the use of algorithmic decision-making systems. (c)

Mitigation of disruption of benefits

The head of an agency that is making a rule under subsection (a) shall, in making such rule, consider whether the use of the algorithmic decision-making system occurs in connection with the administration of a Government service or public benefit, including by a contractor or service provider, and, where practicable, shall seek to mitigate unnecessary adverse effects of the rule on the delivery, accessibility, timeliness, integrity, or continuity of such service or benefit. (d)

Guidance

The Director of the Office of Management and Budget, acting through the Office of Information and Regulatory Affairs and in consultation with the Director of the Office of Science and Technology Policy, shall issue guidance to resolve conflicts and ensure consistency across agencies regarding the rulemaking authority under subsection (a). (e)

Periodic review

(1)

In general

Not later than 3 years after promulgating a rule under subsection (a), and thereafter at intervals determined appropriate by the head of the agency (but not less frequently than once every 5 years), the head of an agency shall review the rule to assess whether, due to technological developments or changes in the use of algorithmic decision-making systems, maintaining the rule is appropriate and that the rule is appropriately tailored to mitigate violations of the Federal law the agency is authorized to enforce. (2)

Repeal and amendment of rules

The head of the agency shall repeal or amend any rule that is determined to be inappropriate to maintain or not appropriately tailored to mitigate violations of the Federal law the agency is authorized to enforce. (f)

Biennial report

(1)

Agencies that engaged in rulemaking

Not later than 2 years after the date of enactment of this Act, and every 2 years thereafter, the head of each agency that has engaged in a rulemaking pursuant to the authority under subsection (a) during the 2-year period preceding the submission of the report shall submit to Congress, the Director of the Office of Management and Budget, and the Director of the Office of Science and Technology Policy, and make publicly available, a report on such rulemaking, including— (A)a description of the rulemaking; (B)the use of an algorithmic decision-making system the agency determined was likely to materially contribute to violations of the Federal law that the agency was authorized to enforce that the agency sought to mitigate through the rulemaking; (C)hiring and staffing actions undertaken by the agency to support the rulemaking; (D)challenges encountered by the agency related to such rulemaking; (E)additional personnel and resources needed to support the use of the rulemaking authority under subsection (a); (F)a summary of enforcement actions taken by agencies pursuant to a rule made using the authority under subsection (a), including— (i)the general categories of violations for which enforcement actions were taken; and (ii)the types and aggregate ranges of penalties or remedies assessed; (G)how the rule will regulate an algorithmic decision-making system to prevent violations of the Federal law that the agency is authorized to enforce; (H)a summary of comments, evidence, or technical assessments received regarding potential disparate impacts or other discriminatory effects associated with the use of the algorithmic decision-making system, and how the agency considered such information in connection with the Federal law the agency is authorized to enforce; (I)in the case of a rule that was repealed or amended pursuant to subsection (e), a summary of the review conducted under such subsection; and (J)a summary of any overlap, inconsistency, or conflict with any other rule or proposed rule of another agency identified during consultation or coordination under subsection (b), and how such overlap, inconsistency, or conflict was addressed or resolved. (2)

Certain agencies that have not engaged in rulemaking

(A)

In general

Not later than 2 years after the date of enactment of this Act, and every 2 years thereafter, the head of each agency described in subparagraph (B) shall submit to Congress, the Director of the Office of Management and Budget, and the Director of the Office of Science and Technology Policy, and make publicly available, a report on— (i)the uses of an algorithmic decision-making system that the agency identified are likely to materially contribute to violations of the Federal laws that the agency is authorized to enforce; and (ii)the factors the agency considered and information on which the agency relied in making the determination not to engage in a rulemaking pursuant to the authority under subsection (a), including any other available mechanisms used by the agency to mitigate the violations of the Federal laws that the agency is authorized to enforce, and a summary of any technical assessment relied upon by the agency in making such determination. (B)

Agency described

An agency is described in this subparagraph if the agency— (i)determined, during the 2-year period preceding the submission of the report, that the use of an algorithmic decision-making system is likely to materially contribute to violations of a Federal law that the agency is authorized to enforce; and (ii)did not engage in a rulemaking pursuant to the authority under subsection (a). (g)

Definitions

In this Act: (1)

Algorithmic decision-making system

The term algorithmic decision-making system means a computational process integrated into software or hardware, including a process based on statistics, machine learning, artificial intelligence, or any other data processing technique, that— (A)is used— (i)to make or execute a decision; or (ii)to assist in making or executing a decision, if such process is capable of altering the outcome of the decision; and (B)is not passive computing infrastructure. (2)

Authority to enforce

The term authority to enforce means, with respect to an agency— (A)authority to issue sanctions (as defined in section 551 of title 5, United States Code) in response to a violation of Federal law or regulation; or (B)the legal authority to bring an action in Federal court to seek a penalty or injunction to remedy a violation of Federal law or regulation. (3)

Agency

The term agency has the meaning given such term in section 551 of title 5, United States Code. (4)

Materially contribute

The term materially contribute means, with respect to a violation of Federal law, to play a significant role in bringing about or enabling such violation, including by assisting or facilitating conduct that results in the violation, but not including minimal or incidental support that is unrelated to the violation. (5)

Artificial intelligence

The term artificial intelligence has the meaning given such term in section 5002 of the National Artificial Intelligence Initiative Act of 2020 (15 U.S.C. 9401). (6)

Passive computing infrastructure

The term passive computing infrastructure means any intermediary technology that does not influence or determine the outcome of a decision, make or aid in a decision (including through evaluations, metrics, or scoring), inform policy implementation, or collect data or observations, including web hosting, domain registration, networking, caching, data storage, or cybersecurity. (h)

Non-Preemption

Nothing in this Act shall be construed to preempt or limit the authority of a State to regulate the use of an algorithmic decision-making system, except to the extent that the regulation of such system by a State is in conflict with this Act or a rule issued under this Act.