New information has emerged in the class action lawsuit against AI company Anthropic which could limit the number of books included in the case—just days before the September 8 hearing in which the details of a settlement among the parties is expected to be released.

Since the lawsuit was announced in July, some experts have estimated that many as seven million books may be part of the lawsuit. But that number now appears to be too high. In a note sent out late Friday, the Association of American Publishers emphasized that the case is limited to books that Anthropic illegally downloaded from the pirate sites Library Genesis and Pirate Library Mirror.

Crucially, the AAP noted that to be included in the class, books had to be published and registered by certain dates. Judge William Alsup’s decision defined an eligible book as “any work possessing an ISBN or ASIN which was registered with the United States Copyright Office within five years of the work’s publication and which was registered with the United States Copyright Office before being downloaded by Anthropic, or within three months of publication.”

“As a bright line rule,” the AAP observed, “any work published or registered after July 2022 is definitively outside the class, and for some works the cutoff is June 2021.” The requirements, the AAP said, could “significantly narrow the universe of works that will be eligible for recovery.” AAP CEO and president Maria Pallante doesn't know that titles will be supplied on September 5, but added that it is her understanding from class counsel the submission made on Friday will include certain information relating to the Works list.

Furthermore, Pallante told PW that from what she understands from the class counsel, that the September 5 submission that the parties are preparing for the court “will begin the settlement approval process, but the Court will provide much more guidance as to next steps, including a notice and claims process, hopefully as soon as September 8,” Pallante told PW.

In another complicating factor, a number of authors shared their realization on social media that some publishers may have not have registered copyright for their books with the U.S. Copyright Office within three months of publication or before Anthropic began its downloading efforts—meaning those books would not be included in the settlement.