I can imagine building a model that is a derived work, but I don't see why all models trained on a corpus that contains some text can automatically be thought of as derived from that text in a creative sense. Models like GPT are using the text as a specific instance of the latent relationships that make up our language. The text is an example of english to learn from, it's not a creative work being extended.

Training a language model, I would argue, is equivalent (in terms of being derivative) to generating a list of the top ten words in a corpus of text. Not really a derived work.

One possible counterpoint- sample based explanation techniques can tell what training data was most influential in a ML model's prediction, and this has been considered in language models[0]. So you could argue that if there are training examples within a corpus that are having an outsized influence on the model output, then maybe it is derivative. This would be pretty cool to look at - are some GPT or other language model outputs relying strongly on a few sources?

[0] https://arxiv.org/abs/1810.03611

It is probably not a derived work. To be a derived work it is not enough to simply use the original work in its construction; it must contain major copyrightable elements of the original work. The weights and parameters present in the model do not seem to fit that description; generally you can’t copyright a bunch of numbers. That said, I don’t think this has been confirmed by a major court case yet, and I wouldn’t be surprised if that happens at some point.

But for a dynamically learning machine model, can you even copyright the numbers?

The question is, "Do the creators of bibliotik have proper licenses to aggregate those books into a dataset?" The dataset is certainly a derived work.

Edit: https://the-eye.eu/public/Books/humble_books_20180509/

Cue the "they're depriving me of my income" complaints from authors.

>The dataset is certainly a derived work

That was not the question.

"Always impressive how the AI community, both inside large companies and not, seems to assume data just exists for them to use, copyright or personal rights be damned."

And you where thinking, lets rephrase that question because it's not already boring and obvious enough?

If AI has human like rights to read and learn and create derived works, what other rights does it have?

Perhaps we should figure that before we start a dystopian nightmare of written material.

Of course it is a derived work.

Maybe not so obvious. If you read a book, are you a derivative work?

Soo reading the Kama Sutra leads to derivative "work" too :)

I read a Java book and created programs using knowledge given in the book. Is my program a derived work?

A book on Java was written to teach you Java, so it would be foolish to claim any copyright on your program.

That opens some interesting questions...

IANAL: I think it's pretty clear it's a derived work, the open question afaik is whether it's fair use our not.