Plaintiffs in Kadrey v. Meta Reference U.S. Copyright Office’s AI Report in Landmark Lawsuit
The plaintiffs in the notable Kadrey v. Meta case have presented a recently released report from the U.S. Copyright Office as evidence in their ongoing copyright infringement litigation against Meta. This report, which discusses the legality of utilizing copyrighted materials for training generative AI models, could greatly impact the case.
Released discreetly last Friday, the Copyright Office’s “pre-publication version” of its final report on artificial intelligence and copyright law questions the assertion that AI companies can invoke the fair use doctrine to justify their data training methodologies. The findings of the report are particularly significant in relation to the Kadrey v. Meta dispute, which features well-known authors and creators including Junot Díaz, Sarah Silverman, and Ta-Nehisi Coates, who claim that Meta used their copyrighted creations without consent to develop its AI models.
Just one day following the release of the report, Shira Perlmutter, the head of the Copyright Office, was dismissed by President Donald Trump. Although the specific reasons for her firing remain ambiguous, some legal analysts and Democratic legislators have speculated that the report may have influenced this decision. Nevertheless, copyright lawyer Aaron Moss suggested in a blog entry that the Office may have expedited the report’s publication prior to any leadership transitions affecting its release.
What the Report Contains
The 113-page document, titled “Copyright and Artificial Intelligence Part 3: Generative AI Training,” marks the conclusion of a series of inquiries into the application of copyright law to AI technologies. It places a strong emphasis on the fair use doctrine—a pivotal topic in the Kadrey v. Meta case—and allocates nearly half of its length to evaluating the four legal factors that determine whether a use qualifies as fair.
While the report refrains from making conclusive legal statements, its findings generally favor copyright holders. It expresses concerns regarding the extensive use of copyrighted materials for training AI systems without authorization, a trend that has gained traction among technology companies.
The plaintiffs presented the report to the court on May 12 as a “Statement of Supplemental Authority.” Their legal filing asserts that the report “addresses several key issues discussed in the parties’ respective motions regarding the use of copyrighted works in the development of generative AI systems and the application of the fair use doctrine.”
Market Harm and the Fourth Fair Use Factor
A key section of the report examines the fourth fair use factor, which considers the effects of a use on the market for the original work. The Copyright Office warns that employing pirated or unauthorized collections of copyrighted works for AI model training could negatively impact the market for those works.
“If thousands of AI-generated romance novels flood the market, fewer of the human-authored romance novels that the AI was trained on are likely to see sales,” the report explains. It also raises alarms about diminished royalty streams and lost licensing avenues for authors—concerns that the plaintiffs have specifically mentioned in their litigation.
Meta, on the other hand, contends that its AI model, Llama, does not compete with the authors’ works and that the model’s outputs are transformative, which it argues bolsters a fair use defense.
Legal and Political Implications
The stance of the Copyright Office contrasts with the interests of leading tech firms, many of which have aligned with the Trump administration’s generally pro-technology regulatory agenda. The release of the report—and the subsequent dismissal of Perlmutter—has drawn attention from legal commentators, though the full ramifications are yet to be realized.
As the report remains in its pre-publication stage, future leadership at the Copyright Office could revise or retract it. Yet, it signifies a notable development in the ongoing legal confrontation regarding the applicability of copyright law to artificial intelligence.
Whether the court will consider the report in the Kadrey v. Meta case is still uncertain. However, for the time being, the plaintiffs are optimistic that it will strengthen their position that Meta’s use of copyrighted materials to train AI systems was illegal.
Disclosure: Ziff Davis, the parent company of Mashable, filed a lawsuit in April against OpenAI, alleging copyright infringement in the training and operation of its AI systems.