Apologies for a second mailing today. This information is of critical importance for writers. Make sure you read it.
What's happening with the Anthropic lawsuit?
There's an ongoing copyright class action lawsuit against Anthropic (the company behind Claude AI) with significant developments recently. Here's the deal.
A federal judge certified a class action suit brought by authors who claim Anthropic violated their copyrights. The key issue isn't that Anthropic used books to train their AI. The court actually ruled that using legally obtained books for AI training is fair use.
Important implications here. U.S. District Judge William Alsup of San Francisco stated in a ruling that Anthropic's use of copyrighted books as training material for Claude was fair use, and the judge found that "using books to train AI models constitutes fair use."
This was a landmark decision that "opens a potential pathway for AI companies to train large language models on copyrighted works without authors' consent — but only if copies of the works were obtained legally."
The ruling was split into two issues:
Legal copies + AI training = Fair use (Anthropic won on this)
Pirated copies = Still copyright infringement (This part of the case continues)
The remaining problem is how they obtained some of those books, not whether they had the right to purloin material from books without the writer's permission.
What remains of the suit alleges that Anthropic downloaded millions of copyrighted books from pirate sites like Library Genesis and Pirate Library Mirror to train Claude. Those are illegal file-sharing sites where copyrighted books are available without permission from authors or publishers.
What does the class certification mean?
The judge approved that second issue as a class action, meaning all authors and publishers whose copyright-registered books were allegedly downloaded from these pirate sites can be part of the lawsuit together. This potentially includes thousands of rights holders.
What's next for authors?
The plaintiffs' lawyers are now reaching out to collect contact information from authors and publishers who might be part of this class. If you have written a published book, you should read that link and possibly register your work for the attorneys. They want to make sure everyone gets official notices about the case and has the chance to either stay in the lawsuit (and potentially receive compensation if they win) or opt out if they prefer.
If you're an author, your books might have been downloaded from these sites for AI training; the lawyers have set up a website where you can confidentially provide your information. You don't need to be certain you're affected—they'll help figure that out.
This case is being closely watched since it could set important precedents for how AI companies can legally obtain training data and what happens when they don't.
Annnd…. appeals are already in motion! Search results reveal interesting details:
The class action lawsuit is "rapidly developing despite an appeal from Anthropic," and Anthropic is seeking to have the 9th Circuit decide whether downloading pirated books to train a chatbot was fair use.
It looks like Anthropic is trying to appeal the part of the ruling they LOST on (the pirated books issue), seeking what's called an interlocutory appeal. Judge Alsup is known to have strong opinions and is not going to change his mind on whether Anthropic's use of pirated material should be condoned. He actually believes the appropriate next step would be a settlement.
This creates an interesting dynamic:
Anthropic won on the fair use ruling for legally obtained books (which they probably don't want to appeal and risk losing)
Anthropic lost on the pirated books issue (which they ARE trying to appeal)
The authors could potentially appeal the fair use ruling that favored Anthropic
Authors' Appeal Options
There is no specific reporting about authors announcing appeal plans yet, but they would logically have substantial grounds and motivation to appeal the fair use ruling that favored Anthropic. The authors have every reason to challenge the fair use ruling because it essentially gives AI companies a legal roadmap to use any copyrighted book for training as long as they obtain it legally (purchase, library, etc.), without needing permission or paying licensing fees. This could fundamentally undermine authors' ability to control and monetize their intellectual property in the AI era.
Given the massive implications for the creative industry, I'd expect the authors' legal team to file appeals, but the timing and strategy may still be developing since this is such recent and complex litigation.
Precedent Impact: The ruling opens a potential pathway for AI companies to train large language models on copyrighted works without authors' consent — but only if copies of the works were obtained legally.
Industry-Wide Implications: The ruling lays the foundations for a precedent that would side with tech companies over creatives.
High Stakes: This is likely only the start of lengthy litigation to resolve one of the most hotly contested questions over the latest tech revolution.
The situation is still developing rapidly, with the ruling not being "a guarantee for how similar cases will proceed, but it lays the foundations for a precedent that would side with tech companies over creatives."
Appeals are definitely happening and likely from both sides—this landmark decision will almost certainly work its way up through the appeals process, especially given its potential to set major precedents for the entire AI and publishing industries.
As a writer, this is what you need to consider:
If you have written a book that was published or self-published, you may be part of the class that evolved from this lawsuit.
There are many other such suits, large and small, developing in the courts.
It's in your best interest to stay on top of these developing suits and the results.
Personal anecdote
I'm a published book author. One of my publishers recently sent me an email advising me that a major player in the AI space has paid a serious sum for permission to use the publisher's backlist to train bots. The publisher gave permission and is allowing its authors to opt in or out of having their work included as part of the AI training. If I opt in, I receive a one-time payment of $ 1000 and grant permission for any of my content published by this house to be scraped and used in AI training models. I feel like there is no guarantee, given the large volume of books, authors, and actors, that if I opt out, my work will actually be protected.
These kinds of situations are going to develop more often as the AI machines become increasingly hungry. Stay aware, writers. Keep your finger on the pulse and take action where you can—your voice is essential.
Again, the place to register as a potential member of the class action is here:
Register your published book(s) with the LibGen & PiLiMi Pirated Books Class —this does not obligate you to take any further steps, nor does it change your legal rights as a class or individual. It's just data gathering at this point.
This is disturbing, Maryan. No doubt there'll be more and more cases like this. The times we live in, I suppose. An important post. Thanks for sharing.
Wow. This is soooooo important. I was surprised by how much the federal court broadened the fair use doctrine in favor of Anthropic's practices. I will look into my published legal books. Thank you, Maryan!