There's such a thing as fair use and digitizing privately owned printed material is absolutely legal... including for corporations.

You get a service. The service is using their compute power to run a model and their scientists to build the model.

Most of their investments fail miserably, but they only need one Google/Stephen King.

Some of us have a good enough memory.

A company can create a new search engine and Google Search isn't obligated to even mention it.

The issue is when achieving market dominance and new service is integrated into the dominant product.

You clearly haven't been around long enough to have caught a lot of discussions on this topic over a decade ago.

I would easily recognize AI generated orchestra, while I can bet a lot of money that AI generated EDM is hard to distinguish from something created in a DAW

They can just come up with a few thousand dollars per hour for a few hours per song. Everyone has a few thousand dollars just lying around

People who like to write lyrics, but can't afford to pay a vocalist still deserve to get their art materialized and distributed.

There's a reason why a few artists persist through the decades, while others just fade into obscurity.(think of how long Madonna, Cher have been around)

AI generated music is also not at all original. Which scares all of the "artists" who lack originality.

It's big business and is really not that deep. For someone who isn't part of that world(which is big business) to judge what is good or not is hard.

And now it's cheap to produce that "pulp romance" novels en masse. So people who have little clue about this genre can produce something that seems good, but doesn't appeal to the reader.

It does not have to be substantial transformation.

I can assure you, that you didn't grant a license with an exclusive list of operations that can be performed on your work of art. At best you may have had something like "no commercial use" clause and general broad terms.

But the problem of attribution is easily understandable to any human with a modicum of intelligence.

Imagine that you have a trillion input images, with every single one having a source associated. When training they go through lots of processes and every single image contributes a varying degree to a subset of 8billion parameters. That alone would produce a dataset that is 1T * 8B to just say how much a particular image contributed to the output...

To mimic intelligence the output is also randomized - the association is not static and every single pixel in the output has it's own lineage.

So as you can probably imagine that to calculate the actual source weights on the output you'd require to do at least 8e+21 calculations per output pixel... and require double precision floating point while you do it.

We know how to do it. It's just ridiculously expensive.

(The above example is for demonstrative purposes only)