But again, even in the case where a painter uses assistants, the limit remains human capability, and a human with the requisite skills did the work. Norms about fair use were established with scenarios like that as context. That still doesn't resemble the paradigm shift that is introduced by generative AI.

It's not about "it should be hard", but rather "artists established norms about fair used based on the fact that it was hard".

Now that this technology exists, those assumptions no longer hold up.