Senators Cantwell, Blackburn, and Heinrich just introduced a bill cleverly titled the Content Origin Protection and Integrity from Edited and Deepfaked Media Act (COPIED ACT).
As they summarize:
- Creates Transparency Standards: Requires the National Institute of Standards and Technology (NIST) to develop guidelines and standards for content provenance information, watermarking and synthetic content detection. These standards will promote transparency to identify if content has been generated or manipulated by AI, as well as where AI content originated. The bill also directs NIST to develop cybersecurity measures to prevent tampering with provenance and watermarking on AI content.
- Puts Journalists, Artists and Musicians in Control of Their Content: Requires providers of AI tools used to generate creative or journalistic content to allow owners of that content to attach provenance information to it and prohibits its removal. The bill prohibits the unauthorized use of content with provenance information to train AI models or generate AI content. These measures give content owners—journalists, newspapers, artists, songwriters, and others—the ability to protect their work and set the terms of use for their content, including compensation.
- Gives Individuals a Right to Sue Violators: Authorizes the Federal Trade Commission (FTC) and state attorneys general to enforce the bill’s requirements. It also gives newspapers, broadcasters, artists, and other content owners the right to bring suit in court against platforms or others who use their content without permission.
- Prohibits Tampering with or Disabling AI Provenance Information: Currently, there is no law that prohibits removing, disabling, or tampering with content provenance information. The bill prohibits anyone, including internet platforms, search engines and social media companies, from interfering with content provenance information in these ways.
The liability provision in this bill seems to make use and copying of content that includes the requisite content provenance information (contemplated by the bill), to train AI models, an actionable offense, with a private right of action. There are no exceptions; no mention of fair use. Section 8 does note: “This Act does not impair or in any way alter the rights of copyright owners under any other applicable law.” But it does not say whether the limitations of the Copyright Act can be impaired.
The COPIED Act applies only to copyrighted content: “The term ‘covered content’’’ means a digital representation, such as text, an image, or audio or video content, of any work of authorship described in section 102 of title 17, United States Code.” Yet, it appears the bill is proposed to exist outside of the copyright system–and apparently not be subject to fair use or other exceptions to copyright, such as idea-expression. If that is not the case, the bill should be redrafted to state so.
