PRIA Logo™
119-s-4915119th CongressWALLET

AI Labeling Act of 2026

Sponsored By: Senator Schatz, Brian [D-HI]

Introduced

Summary

Clear, machine-readable labeling for AI-generated content would be required for generative AI providers and extended to covered online platforms. The bill would set technical standards for provenance, require detection and tamper-resistant labels, and create multiple enforcement paths.

Personalized for You

How does this bill affect your finances?

Personalize government policy and PRIA will tell you what this bill means for your household, plus every other piece of legislation we track. PRIA reads each provision against your financial profile to show you exactly what matters to your wallet.

Free to download

Bill Overview

Analyzed Economic Effects

5 provisions identified: 2 benefits, 0 costs, 3 mixed.

NIST and FTC standards and enforcement

If enacted, the bill would direct the NIST Director to form an "AI-generated content consumer transparency" working group within 90 days. The group would publish technical standards and guidance within one year. NIST would send legislative recommendations to Congress within 180 days after publication. The FTC would be able to treat violations as unfair or deceptive, write rules, and set interoperable standards that, if followed, count as compliance.

Bans and penalties for fake labels

If enacted, the bill would ban knowingly adding false provenance, hiding required AI disclosures, or selling tools made mainly to remove or fake disclosures with intent to deceive. It would also ban knowingly distributing covered AI-generated content without required disclosure for financial benefit. The U.S. Attorney General could sue in federal court, seek injunctions, impound devices involved in violations, and ask courts for damages and remedial orders. The Attorney General could elect statutory damages up to $2,500 per act or up to $25,000 per violation prior to final judgment. Courts could treble awards for repeat offenders within three years and could reduce awards for innocent mistakes. Private plaintiffs and State attorneys general would generally need to notify the Attorney General and the FTC before filing suit.

Labels for AI systems and chatbots

If enacted, anyone offering an AI chatbot across state lines would have to show a clear, conspicuous label that it is an AI chatbot. The bill would define "covered AI-generated content" using a two-part test: the AI must materially change the content's meaning and a reasonable person would think the content was not AI-made. Providers of generative AI would have to embed a machine-readable disclosure saying the content is AI-made, name the system and version, record creation date and time, and include other relevant info consistent with standards. Disclosures would not have to include users' personal data. Providers would also need to make outputs detectable without undue financial burden and help platforms preserve provenance information.

Large platforms must show AI labels

If enacted, a "covered online platform" would be a public-facing site or app that mainly hosts user content and either had at least 10,000,000 monthly U.S. users in the past year or had more than $1.5 billion in revenue in the last taxable year. Those large platforms would have to clearly identify displayed AI-generated content when a machine-readable disclosure is present and would be prohibited from tampering with or removing those disclosures, including when content is shared. Platforms must give users the option to make provenance information available to others and try to follow FTC-recommended strategies to reduce misleading uses. Identification and provenance information should be accessible to the extent technically and economically feasible.

Exemptions for researchers and nonprofits

If enacted, the bill would exempt certain activities from the labeling and prohibition rules. Content generated only for internal research and not intended for public release would be exempt. Nonprofit libraries, archives, and educational institutions would generally be exempt unless they willfully break the rules for commercial advantage; a first offense could bring civil remedies and repeated offenses could forfeit the exemption. Good-faith researchers would be allowed to disable or test disclosures to improve detection tools. Lawfully authorized government investigative, protective, information security, or intelligence activities would also be exempt.

Sponsors & CoSponsors

Sponsor

Schatz, Brian [D-HI]

HI • D

Cosponsors

  • Sen. Curtis, John R. [R-UT]

    UT • R

    Sponsored 6/24/2026

  • Sen. Warner, Mark R. [D-VA]

    VA • D

    Sponsored 6/24/2026

  • Sen. Collins, Susan M. [R-ME]

    ME • R

    Sponsored 7/13/2026

Roll Call Votes

No roll call votes available for this bill.

View on Congress.gov
Back to Legislation