OpenAI now tries to hide that ChatGPT was trained on copyrighted books, including J.K. Rowling’s Harry Potter series::A new research paper laid out ways in which AI developers should try and avoid showing LLMs have been trained on copyrighted material.
I don’t get why this is an issue. Assuming they purchased a legal copy that it was trained on then what’s the problem? Like really. What does it matter that it knows a certain book from cover to cover or is able to imitate art styles etc. That’s exactly what people do too. We’re just not quite as good at it.
A copyright holder has the right to control who has the right to create derivative works based on their copyright. If you want to take someone’s copyright and use it to create something else, you need permission from the copyright holder.
The one major exception is Fair Use. It is unlikely that AI training is a fair use. However this point has not been adjudicated in a court as far as I am aware.
It is not a derivative it is transformative work. Just like human artists “synthesise” art they see around them and make new art, so do LLMs.
Transformative works are not a thing.
If you copy the copyrightable elements of another work, you have created a derivative work. That work needs to be transformative in order to be eligible for its own copyright, but being transformative alone is not enough to make it non-infringing.
There are four fair use factors. Transformativeness is only considered by one of them. That is not enough to make a fair use.
this is so fucking stupid though. almost everyone reads books and/or watches movies, and their speech is developed from that. the way we speak is modeled after characters and dialogue in books. the way we think is often from books. do we track down what percentage of each sentence comes from what book every time we think or talk?
Aye, but I’m thinking the whole notion of copyright is banking on the fact that human beings are inherently lazy and not everyone will start churning out books in the same universe or style. And if they do, it takes quite some time to get the finished product and they just get sued for it. It’s easy, because there’s a single target.
So there’s an extra deterrent to people writing and publishing a new harry potter novel, unaffiliated with the current owner of the copyright. Invest all that time and resources just to be sued? Nah…
Issue with generating stuff with 'puters is that you invest way less time, so the same issue pops up for the copyright owner, they’re just DDoS-ed on their possible attack routes. Will they really sue thousands or hundreds of thoudands of internet randos generating harry potter erotica using a LLM? Would you even know who they are? People can hide money away in Switzerland from entite governments, I’m sure there are ways to hide your identity from a book publisher.
It was never about the content, it’s about the opportunities the technology provides to halt the gears of the system that works to enforce questionable laws. So they’re nipping it in the bud.