AI disqualifies anyone as a musician! It's like playback.

[starflakeprj]

One thing I often remind people talking about Elton John is he didn't write the words of his best.

Neither did Elvis. He hardly wrote anything, but everyone already knew that

[BBFG#]

Which raises the question "is a musician different than a performer?"

I've read often that Michael Jackson would sing his ideas to each musician to actually play the instrument how he wanted it played.
Seems like that would be a better use of AI too.
But I've had a problem with calling AI AI for a long time. Somehow it's always felt like a misnomer to me. And in both words connected, almost an oxymoron. There are more people with "artificial" intelligence than there are computer programs IME. And of course if they're the input, then it's just a case of "garbage in, garbage out." And we've heard countless examples to that in these threads now.

[Bunny_boy]

Which raises the question "is a musician different than a performer?"

I think saying that to an orchestra might elicit an interesting response! :-p

[BBFG#]

Indeed!
And I've met over a thousand of virtuoso level players across many instruments that can't play anything that isn't prewritten for them. And I haven't been able to figure out how that happens...

[Bunny_boy]

Indeed!
And I've met over a thousand of virtuoso level players across many instruments that can't play anything that isn't prewritten for them. And I haven't been able to figure out how that happens...

Have you mentioned that before? If not, I've definitely read (or watched) a discussion about the way classical players are different from jazz / improv players.

[tq]

I find it surprising that the most outspoken enemies of AI in music production have no actual experience with it, or have just heard weird stories about it, or just very briefly prompted a cheap one-liner in Suno and were wondering that it outputs "something". And then they still use modern DAWs, semi-automatic equalizers, samples, loops, superior drummer, orchestral libraries etc. pp. I am old enough to remember the discussions about samples being the death of music, plug ins being the death of 'real' synths, physical modeling to be the death of real instruments, DAWs to be the end of the real performers, and all of this being the end of music.

For the ones that have never used AI some fun facts, and maybe questions to think about...

- You can use something like Suno to 100% stick to your own composition - with an audio file of your song as the input - and just improve the mix/recording. Are you no composer anymore, as you used AI then?

- You could also add some small elements, solos, a background choir, exchange the male vocals for female vocals, even on a pretty granular and controlled level. The composition is still your own - so it's much more like hiring a cover band or a singer. So no "prompt jockey" - would that be cheating? And where's the difference to something like Superior Drummer or using samples, or even hiring (expensive) studio musicians?

- Ace Studio or Synthesizer V or Vocaloid allow to write vocal performances. It's a god send for people who cannot sing, or have no access to a background choir, or want to swap voices. Input can be MIDI, but also your own singing. Would that be cheating? Are you no composer anymore?

- Same for all the mixing tools out there - some of them automatically adapting to your song. Is it cheating - as the only really mastering is somebody turning a knob on a physical compressor or limiter?

- If you let the AI write a song, and then a band performs it, with a real singer - is that performance fake?

I think in all these discussions, we should differentiate between musicians as composers or performers, and between AI as a tool, interpreter/performer, co-writer, editor or composer. These things are not the same, and for some uses, AI is an amazing help - for some uses, it's a cheap replacement of human creativity.

[enroe]

...
I think in all these discussions, we should differentiate between musicians as composers or performers, and between AI as a tool, interpreter/performer, co-writer, editor or composer. These things are not the same, and for some uses, AI is an amazing help - for some uses, it's a cheap replacement of human creativity.

The distinction is of course a prerequisite for thinking in
general – that's implicitly behind the statement: "AI disqualifies
as a musician! It's like 'playback'!" Therefore, I will break this
down a bit:

So ... AI isn't just a "new tool," it's more like a revolution, similar
to the industrial revolution in the 19th century. Nevertheless, many
extremely naive ideas are circulating in this thread. We need to
differentiate:

1. AI tools + plugins
-----------------------------------

There will be AI equalizers that automatically analyze the material
and make suggestions: more open, closer, further away, more
energetic, more mid-focused, etc. All these attributes can then
be selected, and the AI ​​will apply the setting it has learned is correct.

The same will happen with compressors. So there will be
dynamic, very helpful, and intuitively usable compressors.

These AI tools will help the composer, the recording engineer,
and the mixer create "their" perfect mix. AI is extremely helpful here.

2. AI song composition - in 1-3 years
------------------------------------------------------

Here we already see platforms like Suno, Boomy, Aiva, Soundful,
Soundraw, Mubert, Loudly, Beatoven.ai, Udio, Makebestmusic,
Jukebox, and others. Their ability to create entirely new songs is
astonishing. However, the songs are full of minor flaws, don't sound
truly optimal, and often you can even hear the stem extraction
noises that were used.

Can these platforms replace composers, musicians, and mixers?

They can't. AI artifacts are too easily audible, the repertoire is too
small, and the mix is ​​too poor if you have the highest standards.

3. AI song composition - in 4 years and later
---------------------------------------------------------------

Meanwhile, the AI ​​creators have been further developed. Just
as the speech-generating platforms are now mostly quite
perfect and hardly make any mistakes, the new AI music
creators have also become much, much better. They can
now create almost perfect pop, rock, dance, trance, or even
jazz music on demand, and in a diverse, varied, and original
way.

==============================================

And this is precisely where the following applies:

Anyone who uses AI for composing, who doesn't write
or play music themselves, has said goodbye to this world as a
musician.

They make a pact with the devil – perhaps even passing
themselves off as a composer – but are merely a player.

I quote:

You're no longer a musician, no longer an artist — you're just
a consumer! You might as well press the play button on a
CD player or select an MP3 from a long list.

You're no longer a musician!

[ksandvik]

Yes, there's something special if the artist is a human, not an algorithm.

[Tiles]

They make a pact with the devil

Your very obvious aversion to AI is a real obstacle here. Just curious, since I asked and you didn’t answer: do you use any equipment from after the 1960s? If so, then by your own definition, you wouldn’t be a musician.

You’re mixing a few valid observations with a conclusion that doesn’t really hold up.

Yes, AI tools for EQ, compression, and mixing are already here and getting better. But they’re still assistive systems, not autonomous creators. Engineers still make the decisions, just faster.

Current AI music platforms like Suno AI or Udio have improved a lot. The real limitations aren’t just artifacts anymore, but control and structure. Getting something specific and musically coherent over time is still hard. And it does not stop with Udio or Suno. This is just a fraction of the AI possibilities.

Your future projection might happen, but you’re treating it as guaranteed. That’s not settled. And very unlikely.

The biggest problem is your conclusion. You’re again equating “being a musician” with manual effort and struggle, without recognizing that the same is still required when using AI. By that logic, large parts of modern music wouldn’t qualify either. People said the same about synthesizers like the Moog synthesizer, drum machines like the Roland TR-808, and DAWs like Ableton Live. All of them were called “not real music” at some point.

What actually can change with AI is the level of abstraction. You can still use AI like any VST instrument, or you can move from playing and programming to directing, selecting, and shaping. That’s still a creative role, just a different one, and it’s nothing new. Just look at tools like Music Maker. Has Music Maker replaced other methods? Clearly not. For many people, it was an entry point into music. It’s what made them musicians in the first place.

So no, using AI doesn’t “disqualify” someone as a musician. It just changes what kind of musician they are. Like synthesizers and DAWs changed what kind of musicians these people are.

What makes this hard to discuss is that you claim to be leading a discussion, but then simply ignore all arguments, questions, and facts. You’re basically holding a monologue about how bad AI is, just blowing off steam. That would be fine if you at least knew some of the basics. As it stands, you’re doing this without even understanding its possibilities and limitations.

[enroe]

They make a pact with the devil

Your very obvious aversion to AI is a real obstacle here. Just curious, since I asked and you didn’t answer: do you use any equipment from after the 1960s? If so, then by your own definition, you wouldn’t be a musician.

...

No, you're wong here. Perhaps you can read again the whole article.

[Tiles]

No, you're wong here. Perhaps you can read again the whole article.

Of course i am, since i am not your opinion

Perhaps read not only my anwser, but read up the possibilities of AI?

[enroe]

@Tiles: Let me give you a hint: It’s about "distinguishing" and "differentiating."
And to that end, I’m providing three different headings — or areas ...

[tq]

Hey people, ca we keep this discussion civil without personal attacks? enroe makes an argument and defends it, which is perfectly fine. I disagree with him, but I appreciate the time he takes to argue (and so far, this has been astonishingly fine and civil, especially according to KvR standards...) . So let's agree we can disagree?

@enroe, regarding your argument: I think your absolute standpoint feels a bit like back in the day when people tried to make the argument that horses are better than cars. I bet there were good arguments for the horses, but as you said - it's a revolution and you cannot stop it. That fight is over. Cars were successful, and AI will stay with us, too. And to be frank, it was probably already over when people started to develop methods that optimize and adapt algorithms towards data. Even a simple regression could be regarded the seed of ML, so where to start, where to stop here?

As I see it, there are ethical and unethical uses of AI, and you can have an opinion on that - but it's nothing we can stop. So if we cannot stop it, why not use it creatively and in a 'good' way? If you don't, you're just standing at the sidelines and scream at the players on the playing field - but you are not participating in the game.

By the way, full disclosure: I work with ML, LLMs and other things that are labeled 'AI' these days in my professional life. I would dare to say that the work we do there (in an academic context) benefits society, and it's even necessary. And from that, I can see how AI could benefit artists and musicians. But I also see the problematic use and unethical things - so you definitely have a point here. Also, I see why the industry is worried about their business model - but then, one might question that business model also. For decades, the industry was in control of very expensive production tools and the distribution process (and arguably still are, at least partially), and much of it was not related to artists and their interests at all. Actually, one could make the argument that AI also democratizes the access to production tools, and makes it easier for artists (composers or performers) to express themselves.

The problem some of us see with your statement is its absoluteness - "it's all bad' does not help very much, as this will actually make bad uses easier (i.e. to stay in the metaphor - the players on the field do not listen to your screams after a while, and then they just don't care about your opinion anymore, as you are not participating in the game).

[Tiles]

@Tiles: Let me give you a hint: It’s about "distinguishing" and "differentiating."
And to that end, I’m providing three different headings — or areas ...

You continue to evade my arguments, even though I have refuted yours. If you were to address them, you would discover the weaknesses in your argumentation yourself.

[enroe]

Hey people, ca we keep this discussion civil without personal attacks? enroe makes an argument and defends it, which is perfectly fine. I disagree with him, but I appreciate the time he takes to argue (and so far, this has been astonishingly fine and civil, especially according to KvR standards...) . So let's agree we can disagree?

Absolutely, that should go without saying.

@enroe, regarding your argument: I think your absolute standpoint feels a bit like back in the day when people tried to make the argument that horses are better than cars. I bet there were good arguments for the horses, but as you said - it's a revolution and you cannot stop it. That fight is over. Cars were successful, and AI will stay with us, too. And to be frank, it was probably already over when people started to develop methods that optimize and adapt algorithms towards data. Even a simple regression could be regarded the seed of ML, so where to start, where to stop here?

Yes and no. The first concepts for neural networks emerged
as early as the 1940s (Warren McCulloch, Walter Pitts, Donald
Hebb, Karl Lashley). However, only now is the available
computing power sufficient to truly realize neural networks.

In this respect: AI — in the form of neural networks — is here
to stay. It is a bit like "letting the genie out of the bottle": We will
never be able to put the genie back in.

As I see it, there are ethical and unethical uses of AI, and you can have an opinion on that - but it's nothing we can stop. So if we cannot stop it, why not use it creatively and in a 'good' way? If you don't, you're just standing at the sidelines and scream at the players on the playing field - but you are not participating in the game.

Exactly. But since AI will have such a huge impact – on society,
the world of work, on us humans – it is all the more important to
take a close look here.

By the way, full disclosure: I work with ML, LLMs and other things that are labeled 'AI' these days in my professional life. I would dare to say that the work we do there (in an academic context) benefits society, and it's even necessary. And from that, I can see how AI could benefit artists and musicians. But I also see the problematic use and unethical things - so you definitely have a point here. Also, I see why the industry is worried about their business model - but then, one might question that business model also.

You do know that tech giants like Meta, Google, Apple, and
others will be optimizing their "business model" through AI, right?
I'm sure you discuss this in your academic circles as well? But
that's beside the point.

For decades, the industry was in control of very expensive production tools and the distribution process (and arguably still are, at least partially), and much of it was not related to artists and their interests at all. Actually, one could make the argument that AI also democratizes the access to production tools, and makes it easier for artists (composers or performers) to express themselves.

No. I don't see it that way.

The "democratization" already began in the 90s, when Karl
Steinberg (--> Cubase) and Gerhard Lengeling (--> Emagic
Creator, Notator, Logic) created software for the Atari (and Pro
Tools already existed then). From that point on, musicians could
also create complete recordings and professional mixes at
home. AI has absolutely nothing to do with "democratization."

In terms of creativity, AI means "cannibalization."

The problem some of us see with your statement is its absoluteness - "it's all bad' does not help very much, as this will actually make bad uses easier (i.e. to stay in the metaphor - the players on the field do not listen to your screams after a while, and then they just don't care about your opinion anymore, as you are not participating in the game).

Exactly. The absoluteness is a very deliberate choice and
intention.

------ Why? -------

Because whenever AI becomes creative instead of a human,
a red line is crossed.

As long as the AI ​​helps me by adjusting guitar tracks, improving
EQ, or enhancing the mix, everything is fine. Because that's
exactly what it's supposed to do.

However, as soon as the AI ​​generates new, original notes, even
completes entire songs, or even plays a whole vocal part or
guitar solo itself, the red line is crossed. Because in the compo-
sition, in one's own notes, in one's own singing or guitar
solo lies the core of human emotion – all the happiness, the
suffering, life – actually everything. It is here the most
profoundly human, emotional expression!

If AI creates this, then the music at that point is dead, sterile,
worthless. The person who sends out a short text prompt — it
could be the label's or record company's business manager —
is NOT a musician. Nor do they have any connection to a real
musician.

But — sigh — you're right about one thing: It will happen that
tens of thousands will generate AI-generated songs through
simple text prompts. Spotify is already flooded with them, and
in just a few years, 99% of Spotify songs will be AI-generated.

The AI ​​revolution means that for humanity, the "creatives" —
that is, artists, painters, writers, and musicians — will be
pushed into a small niche of insignificance. I find that dramatic,
and that's why I make the very absolute statement:

If you generate a song using an AI prompt, then you
are not a musician. You're just playing back music.