What we should discuss

This note sets aside the legal and ethical disputes over training data for the moment and focuses on how generated work is labeled and where it enters the market. What follows is a personal judgment based on the platforms and communities I have actually spent time in.

Looking around the market

I wanted to hear what other people were making, so I searched through nearly every platform and community I could find, both domestic and overseas. I joined some of those communities, stepped inside the spaces people had built, and left again.

One thing became clear.

AI music discussions are caught in a loop. Is it real music? Is it creation? Should someone using tools such as Suno or Udio be called a musician? At some point, the exchange stops moving.

What I found

Not everyone using generative AI arrived from completely outside music. I play instruments and have performed onstage. More people than I expected had gone much further. It was not difficult to find people with similar backgrounds, working professionals with more than a decade of experience, and others who had spent several decades in the field.

It would be wrong to treat the entire group as musically informed. It would be equally wrong to treat them all as people who know nothing about music.

AI-assisted music-making can matter as experimental work or as a creative practice supported by tools. For that to work, the legal and rights standards need to be clearer, and the process behind the work needs to be labeled honestly. Neither is sufficient at the moment.

Where the problem begins

The people who try to describe their process honestly end up policing every word they use. Meanwhile, people producing so-called "AI slop" push generated work onto platforms without much thought. What remains around them is a market for success stories, prompt secrets, and the promise that anyone can become something without doing the work.

Someone who distributes low-effort output anonymously at scale, presents generated material as something they personally performed, or profits by selling other people a fantasy of effortless success is not the same as someone who discloses the tools and process and uses them as part of a creative practice.

Most of the people I met were using these tools for self-discovery, expression, an expansion of their creative work, or personal satisfaction. They were not trying to claim some grand musical achievement. Some took the result seriously. Some were trying to make something of their own within the limits and conditions of the tools.

The AI market is currently overheated by a rush to secure the first advantage, but that is not a pathology unique to AI. Similar patterns appear whenever a new production technology arrives. Home recording, sampling, YouTube. Each went through its own gold rush. This time the speed of the change and the scale of the market are different, so the fatigue is bound to be heavier.

What could change

There should be a space for different kinds of content made with AI. It has its own room for experimentation and play, and it can become meaningful work. But it should not be mixed without distinction into markets where craft, training, professional experience, authorship, and real-world practice have long formed price and trust.

Clear labels and tags could reduce some of that confusion. A single AI label is not enough. The better questions are more specific.

Was the track generated entirely from a prompt? Did a person restructure, edit, or arrange it? Did it receive a separate mix and master? Does it contain instruments or tracks recorded by a person? Were the vocals generated, recorded, or transformed from a recording? How much of this process has the maker disclosed?

These layers are more useful than the current all-or-nothing distinction. What was made matters, but so do who did what and where responsibility begins.

Without such a system, listeners cannot know what they are hearing, and creators cannot know which market they are entering or under what conditions they are competing. Careless mass production remains on one side, and suspicion toward everyone remains on the other. When the person who discloses the process and the person who hides it receive the same treatment, honesty carries the cost.

So perhaps we can ask something different now. Not whether the work is real or fake, but who did what and how that process should be disclosed. Which market should hold the result, and by what standards? Who carries the rights and responsibilities that follow?

That seems closer to what we should actually be discussing.