The US District Court for the Northern District of California on Monday has given its final approval to a $1.5 billion settlement in the Bartz v. Anthropic case , alleging that Anthropic had used pirated books as material to train its Claude AI platform.

The settlement came after a year in the courts, a judge retiring, and preliminary approval of the settlement in September. Preliminary approval came after a ruling on the training. Prior to the class action lawsuit being certified , Judge William Alsup held that AI companies can use legally purchased books for training, not illegal ones. In commenting on the case, Judge Alsup stated:

“Anthropic had no entitlement to use pirated copies for a central library. Creating a permanent, general-purpose library was not itself a fair use.”

In distinguishing the purchased books from the pirated ones, Judge Alsup summarized that the purchased, printed books being digitized for the central library was a means of saving space. The books themselves were still purchased and it was just a matter of storage. Pirated books, on the other hand, Judge Alsup stated, “could have purchased or otherwise accessed lawfully.” He ruled that the payout will be $3,000 per book covering over 400,000 books.

The monumental June ruling was hailed as a major victory for the AI industry, and that assessment largely still stands. At the same time, the accompanying a $1.5 billion dollar copyright settlement, is the world’s largest copyright settlement to date. The ruling still permits AI companies to train their platforms on legally acquired copyrighted books.

This is the first settlement out of dozens of AI copyright suits still working their way through the courts system. As for the settlement itself, Judge Araceli Martínez-Olguín rejected objections that the settlement price was too low. Quoting the preliminary approval, Judge Martínez-Olguín highlighted that the settlement brings value and closure. Additionally, Anthropic’s attorneys highlighted that over 91 percent of relevant authors claimed their payouts.

The attorneys’ fees portion of the settlement was over $100 million. Judge Martínez-Olguín acknowledged this, remarking that “the class received quality representation from experienced attorneys.” This historic decision signals the end of the unregulated “Wild West” era of AI training data, effectively forcing tech giants to exercise more responsibility in building their models.