What follows is not legal advice. It is merely observations made by reading various sources on the issue.
As many authors have heard, there has been a settlement on a lawsuit over the Anthropic AI company’s use of books to train their AI (artificial intelligence) engine. The understanding is that the books had been pirated by others, but Anthropic used that content. They used 7 million books that had been compiled into two datasets called LibGen and PiLiMi.
The settlement is for $1.5 billion. To be divided among books identified as eligible to be included in the class action lawsuit. Of the 7 million used, only about 500,000 have been identified and classified as eligible for inclusion in the settlement.
Be aware that the settlement is only preliminarily approved. The final approval is expected to occur sometime next year. The earliest the payout will be in June 2026, but this is quite likely to change. Especially if there are appeals lodged against the settlement. Additionally, every claim for each book will need to be verified by the court, which will take some time.
Meanwhile, make a note of the official settlement website. https://www.anthropiccopyrightsettlement.com/
If you are a member of the Author’s Guild, you’ve received very helpful information already, which will be updated. Additionally, Victoria Strauss of Writer’s Beware has an in-depth article on the entire history of this event, providing even more details: Click here for the article.
However, not everyone has taken advantage of the Author’s Guild services or read Victoria Strauss’ extensive article. Thus, I’ll try to provide some basic information and answer a few common questions our agency has already received.
1) Go to the Lookup page on the official settlement website and enter your name (or pen name) to see if your books are included in the settlement database.
https://secure.anthropiccopyrightsettlement.com/lookup/
— If your book or books are not listed, yours did not qualify. You cannot add your book to the settlement or dispute its lack of inclusion. How did they determine which books were improperly used and then eligible for settlement? See question #7 in the FAQ on the settlement website:
https://www.anthropiccopyrightsettlement.com/faq
— Basically, your title had to have been pirated and then included in the dataset that Anthropic used to train their AI. If yours wasn’t pirated, it wasn’t used.
— Additionally, your work must have been officially registered with the copyright office before August 10, 2022, to be eligible. If yours was not officially registered with the copyright office before that date, your book isn’t eligible.
2) If you have books included, it is recommended that you File a Claim using the settlement portal:
https://secure.anthropiccopyrightsettlement.com/
This will take some time and effort to compile if you have a lot of books. Your effort will ensure that your books are included in the final settlement. You have until March 23, 2026 to register.
If you do nothing, you risk losing out on the claim (see question #46 in the settlement’s site FAQ.
3) Getting the money due you
Based on the total announced ($1.5 billion) and the number of books eligible for settlement (around 500,000), it has been reported that each book will receive $3,000. That may be true-ish, but it is unlikely that the entire amount will appear in your checking account. Remember, the court always establishes legal fees, which are paid from the settlement amount. Anyone speculating on the actual amount coming to you is only guessing. No one knows at this time. So, please don’t buy that boat just yet.
This is where it can be murky, and much depends on your particular situation. I cannot answer your questions in the comments with any certainty. I can only suggest some potential outcomes.
a) If your book is traditionally published and still under contract with your publisher, it is possible the money will go to the publisher.
1. The publisher may choose to pass through the entire amount to you.
2. The publisher may choose to treat the income as a subsidiary right (like translation rights, etc). The split between the author and publisher would be determined by the book’s contract. Typically, it is 50/50.
3. Even more likely, since this wasn’t a rights sale, it will be considered a settlement with the rights holder, your publisher. It was a copyright violation. The publisher didn’t sell the rights to Anthropic. The books were pirated. The pirated books were then used improperly, in violation of copyright law. Therefore, if the publisher currently holds the publication rights, they will receive the settlement monies. But then, how it will be split with the author is undetermined at this time. It might be split 50/50, or it might be different. Your mileage may vary if your books are with multiple publishers.
4. Will your agent be paid too? Again, depends on the contract. If your publisher pays 100% of your earnings to the agent, the agency will then split according to your arrangement. If the publisher splits the revenue, sending two checks (one to the agent and one to the author) simultaneously, then that is how it will be handled.
b) If your book was traditionally published but you have the rights back, then the money will come to you directly.
c) If your book was independently published or self-published, it is likely that the money will come to you directly.
Both b and c above will depend on the information you provide when you File a Claim.
To illustrate the complexity of this process, I am aware of a situation where a book was independently published with multiple authors. However, someone included the cover artist when registering the copyright. So the cover designer is showing as a co-author of the book!
In the above situation, with multiple authors, if one of the authors opts out of the lawsuit, which they can legally do, the entire work will be removed from the lawsuit, even if the other co-authors file a claim. (And this means each author in a multiple-authored Indie work will need to file a claim separately.)
Not the Last Lawsuit
This is unlikely to be the last settled lawsuit over A.I. Apparently, there are at least 75 lawsuits worldwide related to the improper use of AI. Click to view the current list of lawsuits as of September 16, 2025, on the “ChatGPT is Eating the World” website.


I suppose that what follows shows that a sonnet can be written about anything, but begs the question… should it?
*****
Lawyers, lawyers everywhere,
by suits they live or die,
but at least they’ve shown they care ’bout piracy for AI.
I guess Anthropic (real weird name)
was developing their machine
that they knew they had to train,
but they didn’t do it clean,
and used a pirated database
of lots and lots of books.
They lawyers went and won the case,
and the billion-dollar payout looks
like penance for the quality
of Artificial Stupidity.
Thanks for this, Steve. I’m fascinated by copyright issues and always love your sharing info on the topic. For me, the most interesting takeaway from this case is the implication of the “piracy” element. Based on the ruling, it seems as though, if Anthropic had merely purchased the 500,000-ish titles individually from legitimate retailers, they could’ve fed them into their AI without violating copyright law. We are in uncharted territory for sure.
Cole,
If Anthropic had purchased each book it isn’t certain if using them as a foundation to train an A.I. engine is proper because it takes the content and “generates” something new. While there have been lawsuits that found “generative” works to be okay…to a certain extent it is yet another case where technology has outpaced the laws.
Thanks for the clarity, brother. Should we let our publishers know when we have filed the claim? Or will they be (have they been) already notified which of their books qualified for the 500,000 titles?
Mesu, I do believe publishers have been notified in some fashion.
We shall all see!
Thanks for posting this, Steve. I found 2 of my earlier books (from the 90’s, one academic, one mass market) listed, and have filed a claim. Will be interesting to see how it goes.
A perfect example of why every author who has published a book should check the database!
Thank you for sharing this! I found 4 of my novels were used to train AI and filled out the form.
My books weren’t on the list, so I’m good. But I’m paying attention because this probably won’t be the last time this kind of thing is done.
Thanks for the info. I muddled my way through the claims process a few days back. 26 of my books were on the list. I found it interesting that most of the ones pirated are still with the various publishers. I’ve had rights returned on over 20 of my older books but only one of those was pirated. And, weirdly, one write-for-hire non-fiction devotional made the list. Obviously I didn’t file a claim on that one. I ran into an obstacle with one book on the list: it’s still listed as being with the publisher, but I got my rights back years ago and it’s currently sitting on CreateSpace under my own name. No idea how that will be resolved. At any rate, I filled out the forms. Now we shall wait and see.
I have eleven books on the list. I filled out the form and submitted it. But according to the info in your post, it seems like I may have made an error. I have the rights back to three of the novels and two of the novellas. But I listed the traditional publisher that was shown on their chart for those books. From Should I have listed myself as the publisher since I have those rights back? Is there a way to correct a form that has been submitted? Or should I submit a new form and change that information?