S2750-119

Introduced

To require the Director of the Office of Science and Technology Policy to establish a Federal regulatory sandbox program for artificial intelligence, and for other purposes.

119th Congress Introduced Sep 10, 2025

Analysis under review: This bill has generated analysis that may be too generic or incomplete. Clause-level evidence remains available below.

Summary

What This Bill Does

The SANDBOX Act (Strengthening Artificial intelligence Normalization and Diffusion By Oversight and eXperimentation Act) creates a federal AI regulatory sandbox program run by the Office of Science and Technology Policy (OSTP). Companies can apply for temporary 2-year waivers or modifications of federal regulations to test AI products and services without being subject to normal enforcement, licensing, or authorization requirements. Waivers can be renewed for up to 4 additional 2-year periods (10 years total). Applications are reviewed by relevant federal agencies with input from private sector and technical experts. If an agency fails to respond within 90 days, the waiver is presumed approved. The program caps at 12 years. The bill also creates a Congressional fast-track process: each year OSTP must submit a report recommending which waived regulations should be permanently amended or repealed, and Congress can act through expedited joint resolutions.

Who Benefits and How

AI companies and startups benefit most directly, gaining the ability to deploy AI products without navigating the full federal regulatory landscape for up to 10 years per waiver. This significantly reduces compliance costs and time-to-market. The program favors companies with US presence (business must be incorporated or have principal place of business in the US). OSTP gains significant new regulatory power as the gatekeeper of the sandbox program. The broader AI industry benefits from reduced regulatory uncertainty and potential permanent deregulation through the Congressional review mechanism.

Who Bears the Burden and How

Federal regulatory agencies bear a compliance burden as they must review sandbox applications within 90 days (with a presumption of approval if they miss the deadline) and potentially lose enforcement authority over sandbox participants. Consumers bear risk because sandbox participants are exempted from regulatory enforcement for the duration of waivers, though existing civil liability is preserved. The FTC potentially loses enforcement power over consumer protection provisions that are waived. Workers in industries disrupted by AI face reduced regulatory protections during sandbox periods. OSTP bears significant new administrative responsibilities.

Key Provisions

  • OSTP must establish the sandbox program within 1 year of enactment
  • 2-year initial waivers, renewable for up to 4 additional 2-year terms
  • 90-day agency review with presumed approval if no response
  • Cap of 120 facilities nationally (5 per state) - NO, this is from another bill. The sandbox has no numerical cap on participants.
  • Consumer disclosure requirements for sandbox participants
  • 72-hour incident reporting for harm, economic damage, or unfair/deceptive practices
  • Existing civil liability and consumer rights of action preserved
  • Director can submit applications on own initiative
  • Annual OSTP report to Congress on all waivers granted
  • Annual Congressional review process with expedited joint resolutions to permanently amend/repeal waived regulations
  • 12-year program sunset
  • State program coordination mechanisms

Evidence Chain:

This summary is generated from the full bill text using AI analysis. Expand "Detailed Analysis" below for identified beneficiaries/burden bearers.

At a Glance

What This Bill Does

Establishes a federal AI Regulatory Sandbox Program administered by OSTP that allows companies to apply for temporary waivers or modifications of federal regulations to test, experiment with, or deploy AI products, services, or development methods. Includes a 12-year program sunset, agency review process, consumer protections, annual Congressional reporting, and a mechanism for Congress to permanently amend or repeal regulations found unnecessary through the sandbox process.

Key Policy Areas

Artificial Intelligence, Government Operations, Regulatory Reform, Consumer Protection

Primary Purpose

Establishes a federal AI Regulatory Sandbox Program administered by OSTP that allows companies to apply for temporary waivers or modifications of federal regulations to test, experiment with, or deploy AI products, services, or development methods. Includes a 12-year program sunset, agency review process, consumer protections, annual Congressional reporting, and a mechanism for Congress to permanently amend or repeal regulations found unnecessary through the sandbox process.

Policy Domains

Artificial Intelligence Government Operations Regulatory Reform Consumer Protection

AI Regulatory Sandbox Program (Title VII, Sections 701-702)

Identified Gains
Contextual inference, no direct clause citation
  • AI Companies and Startups
  • OSTP (Expanded Regulatory Authority)
  • AI Industry (Broader Ecosystem)
Model: N/A | Version: bill_summary_v2 | Source: is

Contextual inference, no direct clause citation

Identified Costs
Contextual inference, no direct clause citation
  • Federal Regulatory Agencies
  • Consumers (Reduced Regulatory Protection)
  • Federal Trade Commission
Model: N/A | Version: bill_summary_v2 | Source: is

Contextual inference, no direct clause citation

Congressional Review of Covered Provisions (Section 703)

Identified Gains
Contextual inference, no direct clause citation
  • AI Industry (Permanent Deregulation Path)
  • Congressional Deregulation Advocates
Model: N/A | Version: bill_summary_v2 | Source: is

Contextual inference, no direct clause citation

Identified Costs
Contextual inference, no direct clause citation
  • Federal Regulatory Agencies (Loss of Authority)
  • Consumer Protection Framework
Model: N/A | Version: bill_summary_v2 | Source: is

Contextual inference, no direct clause citation

Legislative Progress

Introduced
Introduced Committee Passed
Sep 10, 2025

Mr. Cruz introduced the following bill; which was read twice …

Stakeholder Effects

cui bono?

How this legislation distributes effects. Mention counts reflect frequency, not effect magnitude.

Government
7 mentions across 3 clauses
-6 negative

Congress, Federal Regulatory Agencies, Federal Trade Commission

Technology
5 mentions across 4 clauses
+5 positive

AI Companies and Startups, AI Industry (Seeking Permanent Deregulation), Large Incumbent Companies

4/5
sections analyzed
Full impact breakdown

Bill Structure & Actor Mappings

Who is "The Secretary" in each section?

Domains
Artificial Intelligence Regulatory Reform Consumer Protection
Actor Mappings
"the_director"
→ Director of the Office of Science and Technology Policy (OSTP)
"applicable_agency"
→ Head of any federal agency with jurisdiction over waived regulations
Domains
Government Operations Regulatory Reform
Actor Mappings
"congress"
→ Both chambers of Congress
"the_director"
→ Director of the Office of Science and Technology Policy

Key Definitions

Terms defined in this bill

7 terms
"agency" §2a

As defined in 5 USC 551

"applicable agency" §2b

Agency with jurisdiction over the enforcement or implementation of a covered provision for which an applicant seeks a waiver

"artificial intelligence" §2c

As defined in section 5002 of the National AI Initiative Act of 2020 (15 USC 9401)

"covered provision" §2d

Any rule as defined in 5 USC 804(3), including associated guidance, FAQs, bulletins, and derivative material, and any rule required by statute

"health and safety risk" §2e

Risk likely to cause bodily harm, loss of life, or substantial adverse health effect to a human (including before birth)

"risk of economic damage" §2f

Likely to cause tangible, physical harm to property or assets of a consumer

"unfair or deceptive trade practice" §2g

As defined per FTC Act Section 5 and FTC Policy Statements on Deception (1983) and Unfairness (1980)

We use a combination of our own taxonomy and classification in addition to large language models to assess meaning and potential beneficiaries. High confidence means strong textual evidence. Always verify with the original bill text.

Learn more about our methodology