Publishing agreements may allow publishers to sell access to your work for AI training purposes.
Ithaka S+R has started tracking publishers' agreements with AI companies. Review their list to see if your publisher has entered any license agreements.
AI companies use a wide variety of content to train their AI models. Some works are legally obtained and licensed for use. Companies are beginning to license the use of more and more content. However, most of the works used to train models initially were obtained legally through the open web or through pirated sites. If your work is available via the open internet, has been included on a pirate site, or your publisher has signed an agreement with any AI companies your work has most likely already been used to train existing AI models.
A provision within copyright law known as fair use may allow AI companies to use copyrighted works for generative AI training purposes without permission. There are currently many lawsuits underway to determine the legality of using copyrighted works to train generative AI. It is very early days, and final decisions on many of these court cases are years away. Currently, initial rulings have been made in two cases both of which to some degree support the idea that the training may be considered fair use. However, there are additional decisions to be made in both cases and appeals will likely be forthcoming. Additionally, there are a number of other cases still making their way through the courts, so the law if far from settled.
Your contracts will determine what your publisher is allowed to do with your work. It is important to understand what your contract allows you and your publisher to do with your work. Language around subsidiary rights may allow publishers to license the use of your works even for contracts signed prior to the advent of AI. Future contracts will most likely include clauses related to licensing for AI training.
Your contract may also govern how you can use AI in the creation of your work or if you can enter your work into AI models.
If you are concerned about information or a lack of regarding AI in your contracts ask your publisher to clarify or add additional clauses to the contract.
A number of publishers are contacting their authors and offering them the opportunity to opt-in to licensing their works for use in AI training. The decision is personal one that should be made with a full understanding of your rights and how generative AI is using copyrighted works for training.
Bartz v. Anthropic PBC is a class action lawsuit under the Copyright Act brought by authors on behalf of copyright holders against Anthropic PBC, an AI company. The Class of copyright holders — consisting of authors and publishers – claims that Anthropic took books from pirate websites Library Genesis (“LibGen”) and Pirate Library Mirror (“PiLiMi”) without authorization.
The Court certified a LibGen & PiLiMi Pirated Books Class made up of all legal or beneficial owners of the exclusive right to reproduce any ISBN- or ASIN-bearing book that Anthropic copied from the two pirate sites.
Check to see if your book was included in the LibGen database. There is not currently a similar tool for PiLiMi.
If your book is included, there is a potential that you may not be included in the class if you have signed your rights over to the publisher. The copyright holder as of 2022 when the pirated content was downloaded by Anthropic is the eligible party for the settlement.
If you believe you may be a potential member of the class in this lawsuit complete the Author and Publisher Intake form.
For more information on the class action settlement visit the Author's Guild information for authors.