I'm not a native english speaker and what i meant was iconic, not ethnic. Fixed the title - my bad.
This experiment is about helping people explore music of other cultures, in any language.
For example, there is likely no band performing Mongolian throat singing in Yoruba, and we explored how it might've sounded like.
If we're going to improve humanity by being pluralistic and open to understanding other cultures, that must come from examining the conditions and content of that culture itself, not any simulation of them. We can't get closer to people by avoiding talking to them, listening to them, reading their writing, eating their food, going to their temples, etc.
AI helps with none of that but it could fool people into thinking they had done those things.
My statement was made prior to any other comments or the later removal of the term ethnic, but I think my argument is still valid. "Don't do cultural appropriation via LLM."
The fact that the OP was actually talking about Esperanto, of all things, shows how off-base that was.
I'm sure you didn't intend your comment personally, but I'm equally sure that the OP wasn't intending anything nefarious. Pouncing on their use of a single word "ethnic" is just the sort of thing that guideline I quoted asks commenters not to do. If you had paused before pouncing, you could have come up with several other things they might have meant.
You could generate something that you claim sounds like that kind of music, but it's unlikely anyone would believe you or find it interesting.
You see, what makes "ethnic" music interesting is it's shared human history. The use of the term here almost borders on bigotry.