Does music made with AI count as music?
28 Oct 2026
"Producers aren't needed any more." You've probably heard it dozens of times in the last two years. AI makes a song in 10 seconds. You type a sentence and seconds later out comes a mixed song with vocals. And then what? Here we look at the data, the industry's response and the technical facts, then leave the decision to you. Both sides have arguments worth taking seriously.
The question: does an AI track count?
Is this music, or an imitation of music? It sounds philosophical but has practical consequences. Does the sound itself make a track "music", or the intention, effort and choices behind it? If a file sounds like music, does it matter who or what made it?
Data: platforms are flooding with AI
According to Deezer, more than half of the songs uploaded to the platform each day are AI-generated: about 90,000 tracks a day. The share was 44% in April and passed half in July (Source: Deezer, July 2026).
Among so many songs, nobody hears anybody. Streaming offers an unlimited archive and makes being noticed harder than ever.
What the industry says
- It can't stand out. Spotify removed more than 75 million spam tracks in a year and no longer recommends bulk-uploaded songs.
- It can't be copyrighted. According to the US Copyright Office, fully AI-generated music isn't granted copyright.
- It can't earn. Deezer withholds royalties on most AI music it detects; TIDAL pays nothing for fully AI music.
Even the platforms don't want it, at least not fully automated output with no human contribution.
What it can't do technically
- Not club-ready. Often no clean intro and outro to mix, and sections don't sit on 16 or 32 beats.
- You can't touch the detail. Stems don't separate cleanly. Want to change one hi-hat? Regenerate the whole thing.
- The sound gives it away. Metallic vocals, blurry highs, low end that falls apart on a big system.
- It can't do the same thing twice. The same prompt gives a different result each time; you can't develop an idea step by step.
It produces sound. But it doesn't work like a producer.
The stance: "AI will put me out of work"
The fear isn't new. The same was said with every new tool: synthesizer, drum machine, DAW, now AI. None turned out quite like that. Each new tool brought forward those who knew how to use it and helped new genres emerge. House and techno themselves were born on tools once dismissed as "machine music".
But: some just write prompts
Some publish whatever comes out of one sentence as "my track" without learning anything. It's a new version of the ghost producer debate: replacing the ghost producer with a machine doesn't change the story. The problem isn't the tool but misleading listeners about whose work and creativity it is.
Our view: AI can enter production as a tool: generating ideas, sound design experiments, speeding up routine tasks. But the ear deciding direction, character and choices should remain human.
The real issue: tools change, the ear doesn't
Use AI as a tool and you speed up. Surrender to it and you lose your own sound. If you don't know what you want, no tool can know it for you.
A producer's value is choosing the right sound among thousands, hearing why a track doesn't work and improving it step by step. Those skills grow by learning, listening and making.
A separate question for DJs: would you play AI tracks?
Do you know where every track in your library comes from? An unknown track you discovered on a streaming platform could now be fully automated. That's not a disaster, but it raises a DJ's responsibility to be selective: checking how it will sound on a club system, whether its structure suits mixing and whether there's an artist behind it is a new part of good curation.
And a DJ's most valuable offer is a selection no algorithm could make. In an age of automatic output, tracks chosen by a real ear become even more precious.
Try it in the booth
Generate a short loop or track with an AI tool. Listen like a DJ: is the intro mixable, do sections sit on 16 or 32 beats, how would the bass sound on a big system? Then build the same idea from scratch in your DAW with your own choices. Compare, and ask which has more "you" in it.
Learn music first, then choose the tool. Production training at Institute of Sound is 30 hours, one-to-one: from sound design to arrangement, you learn to find your own sound. Would you listen to an AI track? For details visit instituteofsound.co or get in touch.







