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

Bill Number

S. 1396

Origin Chamber

Senate

Congress

119th Congress, Session 1

Policy Area

Science, Technology, Communications

Status

Introduced

Latest Action

2025-04-09: Read twice and referred to the Committee on Commerce, Science, and Transportation.

Last Updated

2025-05-19T15:54:23Z

AI-Generated Summary

Purpose

The legislation aims to promote transparency in digital content by requiring the inclusion and protection of "content provenance information" (machine-readable data tracking the origin and history of digital media like images, videos, audio, or text). It seeks to combat the risks posed by artificial intelligence (AI)-generated or modified content, such as deepfakes (realistic but fake media that misleads viewers), while safeguarding artistic and copyrighted works from unauthorized use in AI training or generation. Overall, it fosters standards, research, and enforcement to build trust in digital media, support innovation, and position the U.S. as a leader in AI development.

Key Provisions

• Standards Development (Section 4): The Under Secretary of Commerce for Standards and Technology (head of the National Institute of Standards and Technology, or NIST) must create a public-private partnership to develop voluntary, consensus-based guidelines and best practices. These cover watermarking (embedding hard-to-remove identifiers in content), detection tools for synthetic (fully AI-generated) or synthetically-modified (AI-altered) content, AI training data transparency, and evaluation methods like "red-teaming" (testing for flaws and risks) and "blue-teaming" (vulnerability assessments and fixes). It includes prizes and challenges coordinated with agencies like DARPA and NSF.

• Research and Education (Section 5): NIST will conduct research on improving detection technologies, watermarking, and cybersecurity to prevent tampering. Within one year of enactment, NIST must launch a public education campaign on synthetic content, deepfakes, watermarking, and provenance, in consultation with the U.S. Copyright Office and Patent and Trademark Office (PTO).

• Requirements for Tools and Platforms (Section 6):
• Starting two years after enactment, commercial tools primarily used to create synthetic/synthetically-modified content or alter copyrighted ("covered") content must enable users to add provenance information and secure it against easy removal (to the extent technically possible).
• It bans knowingly removing, altering, or disabling provenance information for unfair or deceptive commercial purposes. Large platforms (those with $50 million+ revenue or 25 million+ monthly U.S. users) cannot tamper with it, except for limited security research.
• Prohibits using covered content with attached provenance (or from which it was illegally removed) to train AI or generate synthetic content without the copyright owner's express, informed consent and compliance with their terms (including compensation).

• Enforcement (Section 7):
• The Federal Trade Commission (FTC) enforces violations as unfair or deceptive practices under existing FTC authority, with civil penalties.
• State attorneys general can sue on behalf of residents for injunctions, damages, or compliance, with notice to the FTC and options for FTC intervention.
• Private lawsuits: Owners of covered content can sue violators (individuals or platforms) for tampering or unauthorized use, seeking injunctions (court orders to stop), damages, and attorney fees. Statute of limitations is four years from discovery of the violation.

• Other (Sections 1, 2, 8, 9): Cites the bill as the "Content Origin Protection and Integrity from Edited and Deepfaked Media Act of 2025." Expresses Congress's view on AI transparency needs. Preserves existing copyright rights and includes a severability clause (invalid parts don't affect the rest).

Significant Changes to Existing Law

• Introduces mandatory transparency tools for AI content creation software, which previously lacked federal requirements for provenance labeling or anti-tampering measures.
• Expands FTC jurisdiction to treat violations as unfair/deceptive acts, similar to consumer protection rules, without creating a new agency.
• Adds a private right of action for content owners to enforce protections directly in court, beyond traditional copyright infringement claims.
• Builds on the National AI Initiative Act (2020) by defining AI and related terms but focuses specifically on media authenticity, without altering core copyright laws (e.g., under Title 17 of the U.S. Code).

Potential Impacts

• Government Agencies: Increases workload for NIST (research, standards, education), FTC (enforcement), Copyright Office, and PTO (consultations). Could lead to inter-agency collaborations and funding needs for prizes/challenges, potentially enhancing U.S. AI leadership globally.
• Citizens: Improves public ability to identify fake or altered media, reducing misinformation risks (e.g., in elections or news). Creators and everyday users benefit from easier detection of unauthorized AI uses of their work, though smaller users may face compliance burdens if using commercial tools.
• International Relations: By promoting U.S.-led standards, it could influence global AI norms, encouraging international adoption and positioning the U.S. competitively against countries like China in AI ethics and tech development. No direct foreign policy mandates, but it may affect cross-border content platforms.

Main Stakeholders Affected

• Content Creators and Owners: Artists, journalists, publishers, and broadcasters whose works ("covered content") are protected from non-consensual AI training or synthetic replication, with rights to sue for violations.
• AI and Tech Companies: Developers of AI tools must add provenance features; large platforms (e.g., social media, search engines like Meta, Google, YouTube) face tampering bans and enforcement risks.
• Users and Consumers: Benefit from transparent media but may encounter new tool interfaces or labels; smaller entities (under revenue/user thresholds) are exempt from some platform rules.
• Government and Regulators: NIST, FTC, states, and intellectual property offices gain new roles in oversight and enforcement.
• Broader Public: Indirectly affected through reduced deepfake harms, though enforcement may prioritize commercial actors.

Notable Legal, Constitutional, or Political Implications

• Legal: Strengthens consumer protection by leveraging FTC powers without new criminal penalties, focusing on civil remedies. The consent requirement for AI training reinforces copyright principles but could lead to litigation over "informed consent" definitions or technical feasibility of security measures. Private actions empower individuals but risk overlapping lawsuits with state/FTC efforts.
• Constitutional: Likely upholds First Amendment rights by mandating disclosures (transparency) rather than banning content, avoiding prior restraint issues. No direct challenges anticipated, as it targets deceptive commercial practices.
• Political: Bipartisan introduction (by Sens. Cantwell, Blackburn, Heinrich) signals broad support for AI accountability amid rising deepfake concerns. It balances innovation (voluntary standards, prizes) with regulation, potentially influencing future AI bills, but enforcement details may spark debates on tech industry burdens vs. creator protections.

This summary was generated by AI and may contain inaccuracies. Refer to the official source document for the authoritative text.

Sponsor

Sen. Cantwell, Maria [D-WA]

Cosponsors (2)

Sen. Blackburn, Marsha [R-TN], Sen. Heinrich, Martin [D-NM]

Recent Actions

• 2025-04-09: Read twice and referred to the Committee on Commerce, Science, and Transportation.
• 2025-04-09: Introduced in Senate

Bill Versions

• Content Origin Protection and Integrity from Edited and Deepfaked Media Act of 2025 — issued 2025-04-09 — PDF (18 pages)