Long story short, yes. The cost and immorality of having to judge what's "AI" made or "AI assisted" or created with a tool, like a piano, that has AI would be too much and unfeasable. I think it's impossible to legislate, in practice and morally, against AI-produced music. Because AI is a tool. However, what I do have contention with is AI companies mass-producing melodies, patenting them, and issuing copyrights on them, to the point nothing that's created is unique. If AI isn't used to these monopolistic ends, then I don't see how it's any different from using any other tool. Mass-producing copyrights so that everyone else has to pay you isn't the same as producing completed songs.

I agree in principle. The problem is how to determine how much AI is ok to use and how to actually measure it. What's the difference between a musician dreaming up a song and recording it and another musician having an idea and using AI tools to reach a completed song? First, how do we determine what AI-powered tools are acceptable or not?  Second, how do we determine what % of AI is acceptable to use? Third, how do we then evaluate each song to know if it's within that %? Conclusion: It's all infinitely resource-impossibly intensive. It's like wishing we could prevent music creation without a piano. It's too late, illogical, and immoral to do so. What we should focus on is legislation that makes sense, is moral, and is feasible.

Then legislation can be made to make sure these "AI songs" you're referring to are properly labeled like tobacco is, for example. Every pack of cigarettes has multiple warnings, and they're forbidden to publicize, not to mention the extra taxes they also take. These are all ideas that make sense to me to be accordingly applied to "AI music". But trying to solve the issue that you've raised in the other direction seems to me just as irrational and impossible as trying to prevent music created with a piano.