Answers to your questions about AI book theft and copyright compensation
AI companies have trained their large language models on copyrighted books. Was this legal? And if not, what can you, a book’s author, do about it? Can you get paid? This post should answer your questions.
Is it legal to train a large language model on a copyrighted book?
At this point, the legal cases regarding this question have not yet been settled.
For example, US District Judge William Alsup ruled in June that such training was “transformative” because LLM outputs are not just regurgitating the text, so it is covered under fair use. But another case regarding law books said it was not fair use.
However, it is pretty clearly illegal to break encryption on ebooks for the purpose of training models on those books. Alsup ruled that AI companies had violated copyrights by using libraries like LibGen that were built by breaking ebook encryption.
Since most of the AI companies used those illicit libraries, they’re liable for damages.
How can you check if your book was stolen?
Use the tools provided by The Atlantic magazine to check to see if your book stolen and put into a database. You can enter your author name once here and check several illicit libraries, including LibGen and Books3. For example, here’s what comes up when you enter my name as an author:
Can you get compensated for having your book stolen?
One case is on the way to being settled: the case against Anthropic that judge Alsup is presiding over.
You can register as part of the class of authors being compensated here, although that list may have been completed by September 15 so it may be too late. You’ll have to provide information about the books you’ve published that may have been ripped off.
Other cases may be settled in the future. Anthropic agreed to pay $3,000 per book, so it’s probably worth seeing if you can get compensated. Some of that money may go to your publisher, depending on the terms of your publishing contract.
Does your book’s copyright have to be registered for you to get paid?
Yes.
You may have heard that you don’t have to register a book with the copyright office to claim a copyright. This is true. As the copyright office states on its website:
In general, registration is voluntary. Copyright exists from the moment the work is created. You will have to register, however, if you wish to bring a lawsuit for infringement of a U.S. work. See Circular 1, Copyright Basics, section “Copyright Registration.”
Registration is recommended for a number of reasons. Many choose to register their works because they wish to have the facts of their copyright on the public record and have a certificate of registration. Registered works may be eligible for statutory damages and attorney’s fees in successful litigation. Finally, if registration occurs within five years of publication, it is considered prima facie evidence in a court of law. See Circular 1, Copyright Basics, section “Copyright Registration” and Circular 38b, Highlights of Copyright Amendments Contained in the Uruguay Round Agreements Act (URAA), on non-U.S. works.
If your book was published by a traditional publisher, it should have been registered with the copyright office. If not, you’ll likely have to do it yourself.
How can I check if my book is registered for copyright?
You can easily check the registration on the web site of the copyright office here. For example, here’s what I found when I put my name into the public search (note that Forrester, my employer, owns the copyrights on the books I wrote while working there):
I published two books with hybrid publishers. Greenleaf Book Group, hybrid publisher for The Mobile Mind Shift, registered it. Amplify, hybrid publisher of Build a Better Business Book did not, because they don’t provide that service. Don’t assume that your publisher took care of this. You should check.
This matters in the case of the lawsuits now going through the courts. According to the Author’s Guild web site, to assert your rights as a copyright holder in the Anthropic suit:
- The work must have been registered within 3 months of publication, OR
- It must have been registered within 5 years of publication and before the download date of August 10, 2022.
How can you register the copyright for your book, assuming the publisher didn’t?
You can register the book with the copyright office yourself. It’s not difficult.
You’ll need the following:
- Your book’s ISBN number. You can find that on the Amazon page or inside the book on the copyright page.
- A PDF of the content of the book. (You can also mail it, but it’s easier to upload it.)
- A fee of $65 charged to a credit card.
Go to the copyright office’s web site here. Despite the silly warning on the opening page, it worked fine with Chrome on my Macintosh.
Then you’ll have to click to register as a new user.
The site is tedious and poorly designed. The password requirements are ludicrous. But if you follow the instructions to go through page by page, including uploading the PDF of your book, you’ll have filed to register it.
That’s what you’ll need to do to make sure you can get compensated in future lawsuits.