Legislation Watch for AI Developers and Registered Copyright Owners: The Federal CLEAR Act

By R. Scott Weide and Matthew Kramer

On February 10, 2026, Senators Adam Schiff (D-CA) and John Curtis (R-UT) introduced the Copyright Labeling and Ethical AI Reporting Act: the “CLEAR Act.”¹ The bill has received endorsements from a multitude of organizations across the creative industries such as the RIAA, SAG-AFTRA, the Authors Guild, ASCAP, BMI, the National Music Publishers Association, WGA East and West, the Television Academy, and the Recording Academy, among others.²

As introduced, the CLEAR Act would require any person or entity that uses a training dataset in connection with the training or release of a generative AI model to submit a notice to the Register of Copyrights of the Copyright Office containing “a sufficiently detailed summary of each copyrighted work in the training dataset” and the URL for the dataset if publicly available.³ The notice must be filed at least 30 days before the commercial release or use of a new generative AI model, even when the use is restricted internally to an organization.⁴ For models already on the market, the notice must be filed within 30 days after the Copyright Office issues implementing regulations, which must come within 180 days of enactment of the bill. The Register of Copyrights is directed to establish and maintain a publicly available online database of all filed notices.⁵

The bill’s definitions of “artificial intelligence” (defined as an automated system designed to perform a task typically associated with human intelligence or cognitive function) and “generative AI model” (defined as a combination of computer code and numerical values that is designed to use artificial intelligence to generate outputs in the form of expressive material, such as text, images, audio, or video) are relatively broad.⁶ However, reporting is limited to works registered under 17 U.S.C. § 408 or pre-1972 sound recordings protected under 17 U.S.C. § 1401, which limits applicability and should encourage copyright owners to proactively register more works.⁷

The bill provides the owner of a registered copyrighted work utilized but not reported in a dataset the right to bring an action in federal district court and seek (1) civil penalties of $5,000 per instance of failure to submit the required notice, (2) injunctive relief requiring the developer to stop using the undisclosed work until proper notice is filed, and (3) attorney’s fees for prevailing copyright owners. In other words, as proposed, the CLEAR Act may function as a bounty, wherein a copyright owner receives their attorney’s fees and the Copyright Office receives the $5,000 fine.⁸ In this way, it provides a remedy for mere failure to provide the required notice regardless of whether the actual use constituted actionable copyright infringement.

As introduced, the proposed CLEAR Act does not attempt to address or resolve whether the use of copyrighted works in generative AI models constitutes copyright infringement. That issue is primarily determined by whether the use of the copyrighted material constitutes a “fair use” under Section 107 of the Copyright Act, with the Copyright Office opining that “some uses of copyrighted works for generative AI training will qualify as fair use, and some will not,” a position that is being borne out in the courts.⁹

Introduction of the CLEAR Act should prompt copyright owners to register their work and generative AI developers to consider the potential liability associated with using copyrighted works even as training materials. Further, given the potential implementation of the Act and the need to file post-enactment notices, generative AI developers should consider ensuring that all training materials are being fully tracked and logged. Both copyright owners and generative AI developers should be aware of the nascent AI licensing industry, which is increasingly favored in western markets and promoted by forced dataset visibility.¹⁰