PRIA Logo™
119-s-1396119th CongressWALLET

Content Origin Protection and Integrity from Edited and Deepfaked Media Act of 2025

Sponsored By: Senator Cantwell, Maria [D-WA]

Introduced

Summary

This bill would require clear, machine-readable provenance for AI-created or edited media to curb deceptive deepfakes and protect creators. It would set national standards, fund research and education, and restrict using marked content to train AI without consent.

Show full summary
  • Creators and artists: Would require platforms and tools to get express consent and set terms or compensation before using content that carries provenance to train AI. It also clarifies that existing copyright rights remain unchanged.
  • Platforms and AI developers: Would have to adopt NIST-led standards and implement tamper-resistant, machine-readable watermarking and provenance on covered tools starting 2 years after enactment, and would be banned from removing that provenance.
  • Consumers, journalists, and enforcers: Would gain NIST research and public education on detection and provenance and could rely on Federal Trade Commission enforcement, state actions, and private lawsuits with a 4-year discovery-based statute of limitations for certain violations.

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

4 provisions identified: 4 benefits, 0 costs, 0 mixed.

Big platforms must keep provenance

If enacted, a "covered platform" would be any U.S.-available site or app that makes at least $50,000,000 a year or had 25,000,000 monthly users for 3 of the last 12 months. Covered platforms could not remove, alter, or hide provenance so users cannot access it. Platforms are not liable only when removal is necessary, proportionate, and limited to security research. The bill would also ban knowingly removing provenance when done to deceive people in commerce.

Consent required to use content

If enacted, companies could not use covered content that has provenance attached to train AI or make synthetic content for commercial use unless the content owner gives express, informed consent. Users must follow the owner's terms, including payment terms. The rule also covers content where provenance was removed or separated in violation of the bill.

Enforcement and private remedies for provenance

If enacted, the FTC would enforce the bill as an unfair or deceptive practice. State attorneys general could also sue for residents but must usually notify the FTC first. People who own covered content could bring private suits if provenance is tampered with or rules are broken. Courts could order fixes, award damages, and lawyers' fees. Suits must start within 4 years after discovery of the problem.

Standards and tool provenance rules

If enacted, the bill would define key terms like "content provenance," "synthetic content," and "watermarking." It would direct NIST to make voluntary standards, run research, and start a public education campaign within 1 year. Two years after enactment, commercial tools that create or change content would have to let users add machine-readable provenance. If a user opts in, providers must use reasonable security so provenance is hard to remove, when technically feasible.

Sponsors & CoSponsors

Sponsor

Cantwell, Maria [D-WA]

WA • D

Cosponsors

  • Sen. Blackburn, Marsha [R-TN]

    TN • R

    Sponsored 4/9/2025

  • Sen. Heinrich, Martin [D-NM]

    NM • D

    Sponsored 4/9/2025

Roll Call Votes

No roll call votes available for this bill.

View on Congress.gov
Back to Legislation