Is it part of your book contract that your publisher may license your book to the makers of ChatGPT or other large language models or forms of generative AI? If given the choice, should you go along with it?

Those are the questions raised in a recent email from a philosophy professor. He writes:

Last year, I published a book with Cambridge University Press. Now, I have received an email from them requesting that I sign an addendum to my contract agreeing to them licensing my book to providers of generative AI. They note that such technologies “offer opportunities and risks.” I imagine other philosophers have been getting similar requests.

I see some of the potential opportunities here (remuneration; getting my arguments read by other philosophers in the form of supposedly original student essays), but I’m having a harder time thinking through the potential risks—in part because I have little familiarity with the work philosophers are doing on AI. Given this, I was thinking that more-informed Daily Nous readers might be able to weigh in on the ethics and prudence of permitting such licensing.

Discussion welcome, especially from those who’ve faced similar offers, as well as from readers with expertise in legal and ethical matters related to AI, intellectual property, and publication ethics.

UPDATE: Via James Klagge (Virginia Tech), I learned that MIT Press is surveying its authors about this topic. The questions are below. If other publishers are doing this, or are announcing policies, please let us know in the comments. Thanks.

MIT Press Survey Questions:

Do you believe that works you have authored should be used to train generative AI systems? How do publisher practices in this area impact your own choice of publishing partners? 

Please share your perspectives on whether academic publishers like the MIT Press should enter into paid licensing arrangements with AI companies seeking to train on textual works under copyright, and under what conditions. For example, should it be a pre-requisite for any potential LLM partner to provide reliable attribution to your work if that work significantly informs a chatbot’s query response?

If the MIT Press enters into any such licensing deals that have the potential to include your own work(s), would you like to be provided with an “opt in” notification? 

If the MIT Press enters into any such licensing deals that have the potential to include your own work(s), would you expect to be compensated, as per other digital licensing partnerships according to your publishing contract? 

Anything else on LLM training practices and licenses that you would like to share?