observationist
today at 2:18 PM
As of now, you can tell the difference for most AI generated music. There's some where you cannot. There is no Turing Test for taste, and the specific constellation of features that represent your particular interpretation of what things like human, best, goodness, excellence, beauty, and any other label you might apply to abstract qualities will be reproduced at a sufficiently high resolution that you will no longer be able to meaningfully discern between human and AI creations. In a blind test, you will prefer the AI product, and your own perceptions and biases will convince you that the AI generation is actually human, because whatever ineffable abstractions you attribute to "human" quality will be replicated, refined, and exploited.
The very act of recognizing some difference is the tool with which the next generation of outputs is refined, until it's so "good" for any and all particular instances of "good" that human perception is insufficient to differentiate the source.
At some point we're going to have to admit that the distinction based on source is a problem, and perhaps there's a lot of nuance in the context of any particular piece of media such that an arbitrary dismissal of a song, or image, or piece of writing, for the mere reason that AI was used to produce it in whole or in part is missing the point.
If you enjoy a song, your enjoyment is real. If you appreciate beauty, your perception of beauty is real. If you feel deeply about a written text, your feelings are real.
How you perceive things, while not entirely conscious, does involve elements of choice. Make the choice to judge things on meaningful merit, and if the next generation of musicians and artists use AI tools to explore new territory, don't dismiss their art and passion and creations out of hand.
An electric guitar is artificial. People used to make the same sorts of "that's not music" statements people are making now about music and art. Imagine being so twisted up over some arbitrary distinction that you miss out on Jimi Hendrix or BB King, or Joe Satriani, or any of the brilliant musicians that have wrung beauty and soul from "artificial" electronic signals.
There was an interesting study recently which showed that if you put a human-written short story up against an AI-written short story, the AI wins. But if you put an anthology of AI-written short stories up against an anthology of human-written stories, the human-written anthology wins.
I see the same thing in music. I accidentally clicked on a couple of AI albums in YouTube. On a minute-by-minute basis they aren't necessarily bad. But if you keep listening, even though the stream is nominally an hour long, it's the same couple of minutes over and over again, more or less.
In the case of music I could see a coder preferring that for their background noise, but for direct listening for its own sake, once the initial impression wears off there isn't anything left.
I'm not necessarily saying this from an anti-AI position, either. This is just the current reality of the situation. At the moment, AI art has a very flattening effect.
What's more, I spent some time at Suno and tried to get it off the beaten track. I was able to get it to create broken music with chopped up words and instruments that were confused about what they were by trying to make an excessively-interesting combination of genres. It broke before I could get anything really interesting going on musically. Possibly if someone spent a lot of time with the higher-touch music tracking tools they could get something interesting happening but I had enough of the same problems there that I bailed. Even if you try to inject your own inspiration, the AI has a very strong flattening effect.
Text I think you could probably do better with. I have not tried to write fiction but I've done a lot of non-fiction writing with it at work. But no matter how I prompt it, it is always flabby. I can style-shift it away from Default LLM Voice, and it's at least somewhat more concise than that, but I can't get it to be truly concise.
keiferski
today at 2:21 PM
I think most of what you’ve written here only really applies if you listen to music without knowing anything about the musician.
That seems pretty uncommon to me, for most people. The most popular musicians in the world are basically celebrity characters, with the music as a key ingredient, not the only one. Do Taylor Swift fans or Kanye fans or [musician] fans just listen to the music and not follow the person? Pretty unlikely IMO.
I also think it’s an entirely false equivalence to say using electronic instruments are like AI music tools. Very different things – an electric guitar doesn’t play music by itself. It’s still a tool at the end of the day.
neutronicus
today at 3:18 PM
> That seems pretty uncommon to me, for most people. The most popular musicians in the world are basically celebrity characters, with the music as a key ingredient, not the only one. Do Taylor Swift fans or Kanye fans or [musician] fans just listen to the music and not follow the person? Pretty unlikely IMO.
Yes, and AI music generation (like auto-tune before it) enables people to choose their celebrities from a wider pool than "the type of dork who practices guitar for 10000 hours".
horsawlarway
today at 2:50 PM
Frankly, this is also how I mostly listen to music - I set a song or two I like, then I let auto-discovery keep adding to the list.
If I happen to like it, I hit thumbs up/like. Otherwise I ignore it.
I sometimes go through and browse musicians, mainly to see if they have other songs I might like, but generally speaking... it's not high on my list of priorities. Then again, I don't give a shit about the "pop" aspect of music at all. It's mainly background noise I put on while doing something else.
---
As an aside:
> I also think it’s an entirely false equivalence to say using electronic instruments are like AI music tools. Very different things – an electric guitar doesn’t play music by itself. It’s still a tool at the end of the day.
I think this is where it gets weird, and I think you're pretty solidly incorrect here. Samplers and grooveboxes absolutely play music by themselves. I think there's also a weird world where things like "Girl Talk" are somewhat spiritual successors to AI music...
Ex - I definitely love girl talk, and I'm not in any way implying that those albums don't take skill and taste, but he's literally just playing samples of other artists. If that's real music (and I'd argue strongly that it IS real music) then I think I struggle to rule out AI generated songs that are edited by someone (and if you've used this tooling, it still requires lots of editing).
intrikate
today at 3:02 PM
> Samplers and grooveboxes absolutely play music by themselves.
I disagree. You cannot take your Akai MPC out of the box and ask it to play music. You have to load samples, you have to arrange them and you have to instruct it to play. That seems like a far cry to me from "playing themselves." You still have to... write the music.
> Very different things – an electric guitar doesn’t play music by itself.
And a model does not play good music by itself. Only slop if your contribution is nil. Models are more like pianos than parrots.
You're correct that AI will probably end up producing the majority of music no one cares about. Like the muzak you hear on elevators. Or for people that just put on a playlist at work and don't really care much beyond having some background noise.
When you're paying attention, and if you actually care about art and music as human expression, then it will matter. And maybe AI music will still "fool" someone then. Maybe we'll discover the next Michael Jackson was just prompting their way to the top of the charts. But that won't really be the point, just like it wasn't the point when everyone discovered that Milli Vanilli were faking it.
People don't like liars. And using AI to generate art is lying. You didn't make it, the AI that did make it was only possible because it collectively stole from every human musician and artist before it. You can wrap it up however you like, but at the end of the day its just a lie.
And yeah there's some nuance here. Lets take Milli Vanilli for example. They were considered frauds because they weren't actually singing on their tracks. What if they had been singing, but using autotune? I don't know where you draw the line but for me its somewhere around people who have no appreciation of the amount of effort that goes into producing art that think they can create it whole cloth from a couple of prompts.