The plaintiff in Landmark Kadrey v. Meta has filed a U.S. copyrighted AI report as evidence of copyright infringement lawsuits against the technology giant.
Last Friday, the Copyright Office quietly released a “pre-publication version” about using copyright-protected projects to train generation of AI models. The follow-up report contained bad news for AI companies, hoping to defend the claim in court that legal use of legal doctrines.
Less than a day after the report was published, Shira Perlmutter, head of the copyright office, was fired by President Donald Trump. It is not clear why Perlmutter was fired, but the move shocked some copyright lawyers, as Mashable previously reported.
On May 12, Kadrey v. Meta’s plaintiffs, including artists and authors such as Junot Diaz, Sarah Silverman and Ta-Nehisi Coates, filed the report in their class action lawsuit.
What is the AI report from the US Copyright Office?
The office’s report is the conclusion of a three-part investigation into copyright law and artificial intelligence, called the unknown legal domain. The “Copyright and Artificial Intelligence Part 3: Generated AI Training” report accurately examines the types of legal issues in Kadrey v Meta.
While some copyright lawyers and Democratic politicians speculate that the report led to Perlmutter’s sack, there are other possible explanations. Copyright attorney Aaron Moss said in a blog post: “Before the leadership wave changes, the office is more likely to announce the conclusions of the report.”
The report details four factors of the theory of rational use. Meta and other AI companies are being sued for using copyrighted engineering to train their AI models, especially Meta claims that the activity should be protected under fair use.
Mixable light speed
The lengthy 113-page report costs about 50 pages, detailing the nuances of fair use, citing historic legal cases to rule and oppose fair use. This does not draw any package conclusions, but its analysis is usually more than AI companies and their unprecedented databases for inventory related to model training.
The Copyright Office’s position on hot white issues is inconsistent with the will of the big tech giants who work towards the Trump administration. In general, President Trump has adopted professional and technical AI regulations.
The plaintiffs in Kadrey v. Meta obviously hope the report will be in their favor. Attorneys who filed the report as evidence on Monday did not explain in detail why it was submitted as a “supplementary agency statement.” The abstract simply states: “The report addresses several key issues discussed by the parties in various motions to develop copyrighted works in AI systems and adopts the application of fair use doctrine.”
Copyright Office finds AI models could hurt creative markets
Controversial AI copyright reports may make case prompts for Meta.
Image source: Wildpixel / Istock / Getty Images
The most outrageous part of the report is the Office of Copyrights’ assessment of the fourth factor of fair use that takes into account the impact on current or future markets.
“Using pirated copyrighted works to build a training library, or distribute such libraries to the public, will harm the market access to these works,” the pre-public version of the report said.
This analysis also considers the author’s possible market dilutions. “If thousands of AI-generated romance novels are put on the market, there is little training in AI’s AI novels. In addition, the plaintiff argues that Meta’s use of piracy to access authors’ books deprives them of their licensing opportunity.
Meta believes that its AI model Llama does not compete with the author market, and that the transformative output of the model makes it irrelevant to its rational use argument.
Although the report favors the plaintiff’s argument, we don’t know if the judge in the case would agree. And because this is a pre-publication version, future leaders of the Copyright Office can edit or revoke it.
Disclosure: Mashable’s parent company Ziff Davis filed a lawsuit against Openai in April, accusing it of infringing on Ziff Davis’ copyright in training and operating its AI systems.
