Anthropic just settled the largest copyright case in American history for $1.5 billion, and the authors who sued say they still didn't get nearly enough.

A federal judge in San Francisco approved the deal on Monday, closing out a class action lawsuit filed in 2024 by a group of writers who accused the AI company of using pirated versions of their books to train Claude, its AI chatbot. The settlement covers more than 480,000 works, and authors filed claims on over 92% of them.

How Anthropic got here

The case had a complicated path to this point. A now-retired judge ruled last June that Anthropic made legitimate "fair use" of the authors' work for training purposes, which is a legal doctrine that allows limited copying without permission. But that same judge found that Anthropic had separately saved more than 7 million pirated books to what the court called a "central library," copies that weren't strictly necessary for AI training. That distinction mattered enormously. It shifted the question from whether Anthropic owed anything to how much, with potential damages estimated in the hundreds of billions of dollars.

Facing that exposure, Anthropic settled.

What $1.5 billion actually means

The lead attorney for the authors, Justin Nelson, called it "the largest known copyright recovery in history." The judge awarded the plaintiffs' lawyers just over $101 million of the $187.5 million they requested in fees.

For individual authors, the math is less impressive. Splitting the remaining sum across hundreds of thousands of works means most writers will receive a modest payment at best, not a life-changing one. Some objected to the settlement precisely for this reason, arguing it was too small given the scale of what they alleged Anthropic took. The judge overruled those objections, writing that the complaints were "not grounded in a realistic assessment of the overall risks and rewards of a trial."

That framing is worth sitting with. Even with potential damages in the hundreds of billions, authors faced a real possibility of losing more at trial, or winning far less than the theoretical ceiling. Settlements like this trade certainty for size.

Not everyone accepted the trade. Some authors and publishers opted out entirely and have filed separate lawsuits against Anthropic that are still working through the courts.

The bigger pattern

This settlement is one piece of a much larger legal fight reshaping how AI companies are allowed to operate. Dozens of similar cases are moving through US courts, brought by news organizations, photographers, musicians, and other creators who argue that tech companies consumed their work without permission or payment to build products now worth hundreds of billions of dollars.

Anthropic, backed by Amazon and Alphabet, is one of the most valuable AI companies in the world. The $1.5 billion it agreed to pay is real money, but it is a fraction of what the plaintiffs originally sought, and a fraction of the company's perceived value.

The more consequential question this case leaves open is what it signals to the rest of the industry. This is the first major US copyright case involving AI training to reach a final settlement. That makes it a data point every other defendant and every other plaintiff's attorney will study carefully. If $1.5 billion is the price for storing 7 million pirated books and building a product as widely used as Claude, that number will land very differently depending on which side of the courtroom you're sitting on.

For creators watching this unfold, the lesson is probably less about the dollar figure and more about the underlying precedent: courts are willing to find that AI companies crossed a legal line, and settlements, not trials, are likely to be how this era of disputes ends.