So the question of whether training #AI on other peoples' work is legally OK or not is an important question.
But. Assume we have generative #AI systems trained only on correctly sourced, legally approved materials.
When and where is it OK to use that output? Mandatory labelling of sources? What threshold to count as "AI"? Should the rules extend to all art (collages, for instance) not just AI generated?
Those questions are far more interesting and important in the long run.
I'm sure it all varies widely by context and use case.
Very often it really doesn't matter one bit where some content came from; its value is in what it is, not from its source.
So it's really probably between individual producers and consumers of the media, to be sorted out by reading the wishes and applications.
