Humans are punished for plagiarism all the time. Myriad examples exist of students being disenrolled from college, professionals being fired, and personal reputations tarnished forever.

When a LLM is trained on copyright works and regurgitates these works verbatim without consent or compensation, and then sells the result for profit, there is currently no negative impact for the company selling the LLM service.

false equivalence because machines are not human beings

a lossy compression algorithm is not "inspired" when it is fed copyrighted input

> lossy compression algorithm is not "inspired" when it is fed copyrighted input

That's exactly what happens when you read. Copyrighted input fed straight into your brain, a lossy storage and processing machine.

I think it’s a pretty easy principle that machines are not people and people learning should be treated differently than machines learning

You see this principle in privacy laws too.

I can be in a room looking at something with my eyeballs and listening with my ears perfectly legally... But it would not be legal if I replaced myself with a humanoid mannequin with a video camera for a head.

You can even write down what you are looking at and listening to, although in some cases, dissemination of, e.g. verbatim copies in your writing could be considered copying.

But it is automatically copying if you use a copier.

Following your analogy, parrots should be considered human.

Issue to me is that I or someone else bought those books. Or in case of local libraries the authors got money for my borrowing copy.

And I can not copy paste myself to discuss with thousands or millions of users at time.

To me clear solution is to make some large payment to each author of material used in traing per training of model say 10k to 100k range.

If your book reproduces something 95% verbatim, you won't even be able to publish it.

Exactly. We assess plagiarism by checking the output (the book), not the input (how many book I’ve read before). It’s not an issue to train LLM on copyrighted resources if their output is randomized enough.

If you are plagiarizing, “for free” doesn’t even save you.