The biggest compensation award in the history of AI copyright has been born! 😱

Anthropic actually has to pay $1.5 billion in a class-action settlement—because when training Claude, it used millions of pirated books without permission!
Although the court found that training counts as fair use, storing pirated books is infringement.
Training ≠ innocence: the court ruled that using data to “train” a large model is fair use, but your server’s act of “storing” copyrighted works is directly classified as infringement.
A $1.5 billion settlement—an eye-watering figure—sets a precedent for re-evaluating the data costs of large models.
This is only a class-action lawsuit. Some publishers and well-known authors have already withdrawn from the agreement and are preparing to sue separately and fight it head-on.
Can’t AI companies be made to pay all the way to the end? After this, it’s likely that a whole bunch of people will sue, too? 😂