AI generated music barred from Bandcamp

Same for me! Switched to Bandcamp + Navidrome and have decided that one of my goals for the year is to find at least 2 albums per month I want to buy.

I will shamelessly promote the bandcampsync [1] CLI tool for automating downloads of your bandcamp library and bandcamp-sync-flask [2] wrapper that I built so I could invoke it from the web on my phone after I buy an album.

[1] https://github.com/meeb/bandcampsync

[2] https://github.com/subdavis/bandcamp-sync-flask

I’ve been doing the same over the last few months.

The best part for me is going to record stores again. CDs are SO cheap now, especially used ones. I’ll usually pick a few out of the dollar bin just based on vibes and the cover and rip them when I get home. I’ve found some cool stuff. It’s like a treasure hunt.

Don’t miss Spotify one bit.

I just wanted to say, thanks for saying this. I actually have been writing music and using Distrokid to publish to the normal streaming services (Spotify/Apple Music) and your comment actually pushed me to sign up for and put music on Bandcamp.

In case you'd like to take a look, shameless self promotion: https://aaronholbrook.bandcamp.com/music

I have to go through my back catalog, and add all my music, but I appreciate your perspective and for wanting to support artists!

I've been reading about Spotify pushing generated music but haven't seen that myself so I'm interested to know in what context it happens. Is it certain music styles? Spotify's own playlists? That smart shuffle feature?

I listen mostly in the old school way, full albums of my favourite artists, so I suppose it would be quite unexpected to stumble into AI music this way.

I switched to Bandcamp a while back because I was sick of Spotify playing the same 100 songs forever. It feels like they have about 2 songs for every artist that they will actually play in any generated playlist.

Best part of owning music is importing them to an old version of iTunes and sync'ing to your iPod.

I am vindicated in my choice to use an ipod with an aux jack every time Android Auto can't decide whether to connect over USB or bluetooth and just doesn't play audio until I restart the phone and the car...

I miss what.cd, bandcamp almost a replacement

Bandcamp continues to be the best place to organically discover new artists. If I'm ever bored I go to their front page and browse by genre. It feels like the digital version of Sam Goody or whichever 90s record store had the headphone kiosks where you could listen to songs before buying the record.

Spotify, on the other hand, induced a level of visceral disgust I'd never felt before when I stumbled across an AI-generated album supposedly made by an artist I enjoy. In this case it was somebody that had been dead for 15 years - they were hijacking her Spotify page to promote it as a new release. I'm not an AI reactionary but I found this absolutely fucking gross. Having AI-generated music for four-hour YouTube videos of anime girls sitting in apartments on a rainy day is fine. Desecrating the body of work of a departed musician is decidedly not.

Same. I buy music from Bandcamp and Qobuz. I don't stream it though instead opting to sync my massive music collection through Syncthing

If you came across a song and fell in love with it, only to find out later that it was generated by ai, would you stop loving the song?

I experienced a similar renaissance! Now using Nextcloud and Subsonic to stream music to my phone etc. Wrote a little article on my blog: https://blog.ture.dev/posts/goodbye-spotify-and-yt-music/

Same.

I like the idea of my money going to the artists. And, you can "buy the catalog", give an artist $150 or so to get ALL their music. I have a couple composers who I adore so that was a no-brainer. If I was going to pay them for most of their work anyway, why not give them the money now?

For those who miss Spotify connect capabilities and apps, I've been using Roon (self-hosted, but susbcription required) for my music library for a couple years now and it is absolutely excellent. You get full access to stream your own library and the ability to integrate into Tidal and/or Qobuz for any music you want to hear but don't own a copy of. It's really very good.

When I used to play, Bandcamp was always the most band/creator-friendly platform

Great to see this! The flood of AI Music will bring music royalties down to zero which Im all for if it kills AI music(spammers and silly prompt engineer songwriters trying to make a buck go away)! AI trying to mimic humanity all ways needs to die! Who does AI benefit besides those at the very top?

*Note i am hobbyist songwriter (melody and lyrics) since a teen (few decades ago)and use Suno. It makes my songs sound just like everyone elses cookie cutter crap... it has no soul to it.. just the feel of tech billionaires getting filthy rich off destroying society/humanity!

Absolutely! Bandcamp has really been phenomenal these past few years for me as well

As someone who's always bought music rather than getting a subscription, welcome to the club!

Do make sure to back everything up, though, I remember when Google Play Music was shut down and I needed to download everything (fortunately it was announced well beforehand so there was no need to rush).

7digital is also pretty good, I've bought a bunch of Saxon and Rainbow albums on there. As awesome as Bandcamp is, many bigger artists don't have a presence there (although King Diamond's entire discography is on there, that's cool).

What do you search for or listen to that gives you ai generated music?

Now that's a sales pitch! You have my interest.

my favourite part of having my own local library again is turning on shuffle and having it actually shuffle

There's something genuinely satisfying about owning the music again instead of just passively streaming whatever an algorithm decides to surface

It took one obvious, obnoxious, and well infiltrated new music AI slop in the my recommended new music feed to finally turn me completely sour on Spotify. I was a prelaunch US user who had brand loyalty built in from the start. I even met Daniel Ek during the big early hype. Its gone. It has been the listen of last resort for awhile, and I used it for discovery of new releases. It's dead to me now

Amen!

I switched to buying CDs and vinyl again a couple of years ago, and bought some second-hand hifi equipment.

I still have a family AppleMusic subscription mostly because my kids use it a lot especially in the car, but I want to go back to owning creative works and compensating artists instead of renting.

Also a musician and I don't think it's that amusing. IMO this isn't an "AI can't be art" discussion. It's about the fact that AI can be used to extract value from other artists' work without consent, and then out-compete them on volume by flooding the marketplace.

AI takes all of that old school idm and electronic music and repackages it without a human story to tell, ripping off actual musicians in the process. AI didn’t magically ‘make old IDM its own.’ It scraped decades of artists’ work, stripped out the context and intent, and reassembled the surface features. There’s no human arc, no lived constraint, no risk and no culture.

What’s being repackaged isn’t a new instrument, it’s other people’s careers. I’m not sure what part of that is supposed to be amusing.

The issue is not so much an artist that will use it as a tool, even though there is much to say about it, it's the hundred of thousands of people with no interest in music whatsoever, that will flood the platforms in order to make a quick buck.

The main differentiator I've noticed is: how much work is the tool doing, and how much work is the artist doing? And that's not to say that strictly more effort on the part of the artist is a good thing, it just has to be a notable amount to, IMHO, be an interesting thing.

This is the primary failure of all of the AI creative tooling, not even necessarily that it does too much, but that the effort of the artist doesn't correlate to good output. Sometimes you can get something usable in 1 or 2 prompts, and it almost feels like magic/cheating. Other times you spend tons of time going over prompts repeatedly trying to get it to do something, and are never successful.

Any other toolset I can become familiar and better equipped to use. AI-based tools are uniquely unpredictable and so I haven't really found any places beyond base concepting work where I'm comfortable making them a permanent component.

And more generally, to your nod that some day artists will use AI: I mean, it's not impossible. That being said, as an artist, I'm not comfortable chaining my output to anything as liquid and ever-changing and unreliable as anything currently out there. I don't want to put myself in a situation where my ability to create hinges on paying a digital landlord for access to a product that can change at any time. I got out of Adobe for the same reason: I was sick of having my workflows frustrated by arbitrary changes to the tooling I didn't ask for, while actual issues went unsolved for years.

Edit: I would also add the caveat that, the more work the tool does, the less room the artist has to actually be creative. That's my main beef with AI imagery: it literally all looks the same. I can clock AI stuff incredibly well because it has a lot of the same characteristics: things are too shiny is weirdly the biggest giveaway, I'm not sure why AI's think everything is wet at all times, but it's very consistent. It also over-populates scenes; more shit in the frame isn't necessarily a good thing that contributes to a work, and AI has no concept at all of negative space. And if a human artist has no space to be creative in the tool... well they're going to struggle pretty hard to have any kind of recognizable style.

Cf. Holly Herndon's album Proto.

This is something people spent a lot of time on, is trained lovingly on only their own stuff, and makes for some great music.

It's "AI" but in an almost unrecognizable way to us now: its not attached to some product, and its not about doing special prompting. It is definitely pop/electronic music, but it follows from a tradition of experimentation between what we can control and what we can't, which is here their bespoke stochastic program.

https://youtu.be/sc9OjL6Mjqo

It is not about how the computer or the model enables us, which is so silly. (As if art is simply about being able to do something or not!) Its about doing something with the pieces you have that only those pieces can do.

AI as tool is included inside almost every daw (or can be through VST) and there is no way bandcamp could enforce a strict "non AI has been used in the process" policy. I think it is sane to separate cases where a record is entirely generated by a single prompt vs AI used as instruments/tools.

I don’t think this is an AI issue, but the amount of effort, the thought process and the story telling about the track they made.

Before generative AI, there were already a swarm of people who aimed at maximising the number of track they made within a short time, with abusing marketing. It is not wrong that they can pump up 100 tracks in a year, with a template and a specialised workflow and correct marketing techniques but… what is the story to these music? For many tracks, I only heard the story of:

> I am the most productive person and I can make most of the money because of that.

Quantity wise, for sure, they wins, but quality wise, I failed to imagine a more complex story than things above although they are good to hype the dance floor or a concert. These days, I mostly listen to music I have bought, or made by specific music communities because of their story behind their track despite not as perfect.

Same reasons why don’t I watch many movies since Ironman 3, most of the blockbusters follow the same winning formula rather than trying something new and in depth or unexpected, CGI and product placements all over the place instead of a good story.

AI just emphasised this problem even more since commercial “art” has been testing majorities’ newest lows.

There are differences between using a tool to create art or use it to spam.

> Even if its just creative prompting, or perhaps custom trained models, someday someone will come along and make a genuine artistic viable piece of work using ai.

We've now had this technology for 2 years. Show me one, just ONE track that is purely(!) made by AI you find honestly exciting. Not "commercially successful", mind you, something you, a musician, personally think is actually great. You are referencing Aphex Twin there, and I'm old enough to remember when I first heard "Digeridoo", so, you know, something where you just go "Wow, that's a banger". If you're DJing: something you would actually put on in a club and the crowd would go wild.

Let's cut the crap: there is none. All GenAI is good for is generating stupid memes, shitposting, ads, and generic background music. There is ZERO creative value in purely generative AI. Yes, there are tools leveraging AI models which can help musicians create tracks - entirely different thing. This is also not what Bandcamp is banning here. Most people will freely admit that AI tooling can be used creatively, like what De Staat did with the "Running backwards into the future" music video - that's all fine, really nobody is disputing that fact, although that "look" is now well established and people are mostly bored and annoyed by it, but that's just how it goes.

There are plenty of places to publish AI generated music. Why should a platform allow music it's users clearly don't want.

I find it interesting that there's so much pushback against ai generated art and music while there seems to be very little for ai generated code.

> demonized until some group of artists came along and made it there own

I'm pretty sure the people at Bandcamp agree with you and that's why they mention future "updates to the policy as the rapidly changing generative AI space develops".

Well that's the issue. We're not seeing "artists" coming along and applying it to their years/decades of creative knowledge. We're seeing the equivalent of some cushy heir to a fortune come in with a drill and say "I can outdo these teams of diggers! We don't need diggers anymore!"

And on the surface the drill is better. But this heir is assuming that all diggers do is diplace dirt. Not thinking about where to dig, how to dig safely,, what to dig for, and where brute force is needed vs a subtle touch (because even in 2025, miners keep shovels with them). That's all going out the window for "hey I made a hole, mission accomplished!".

Instead of working with diggers to enhace their mining, they want to pretend they can dig themselves. That's why no one in the creative space is confident in this.

> its just creative prompting,

Sure, you just can't upload the resulting track directly on Bandcamp, but you're free to "creatively prompt" on SUNO all you want, they'll even host your "music".

It's also a matter of resources. People uploading gigabites of AI generated slop a day isn't really what Bandcamp is about.

"greetings, fellow musicians" - genAI and quant guy

I’m more hopeful that MIDI completion/in-filling models will be easier for musicians to control and use. But right now, the most popular tools are things like Suno, where you barely have any control and it spits out an entire, possibly mediocre song. It’s the same vein as ChatGPT image generation vs. Stable Diffusion, where you can do much more controllable inpaints with the latter.

>someday someone will come along and make a genuine artistic viable piece of work using ai

And in the mean time, AI will continue to clutter creative spaces and drown out actual hardworking artists, and people like you will co-opt what it means to be an artist by using tools that were trained on their work without consent.

It's like the reverse of the product that advertises itself as "AI driven." As if that's supposed to be a selling point. OK, it's AI driven, but is it good?

There may be short term emotional strings to pull. "AI driven!" or "AI free!"

But ultimately, no one will care if it's AI or not if it's good.

People don't listen to music because it sounds good, most music sounds downright awful, they listen to it for the human stories and connection. 99% of the music I listen

Nobody listens to techno - Eminem

AI needs to make music that sells. The same way Scorsese and Brando sell the Godfather. until it can do that literally nobody will care.

You can already do this, but the platformised generative AI is sloppy by comparison and not that interesting.

https://github.com/acids-ircam/RAVE

There is a difference between Richard D. James hand-training an LLM on foley sounds he recorded himself to put in his latest IDM track, and the script kiddie spamming out 50 AI-generated mixes per day to get that sweet ad revenue on Youtube.

No one is complaining about the first case, because they are outnumbered by the second 100,000 to 1. RDJ isn't gonna use suno.ai no matter how pro-LLM he is.

Note: this is for sake of argument, I am not aware of RDJ using LLMs in any shape or form.

So you just want to be lazy and subsidize to the parrot machine the very essence of what it means to be creative. I am utterly baffled by this recurring comparison between past electronic tools, which actually have a pretty harsh learning curve to be mastered, and a software contraption that overtakes your creative agency. I see it everywhere, like comparing Midjourney to the shift to digital photography. What are y’all blokes on? How is it possible that even fine minds just lazily accept such a flawed parallel between two completely different technological paradigms?

This is not a super well thought out position, but I've been leaning towards really disliking AI art in general (without having an opinion on any strong policy action yet).

First, art is, I think, one of the most enjoyable activities we have. One evidence is a lot of people forego higher salaries to choose an art job (although being a job carries additional responsibilities and some inconveniences compared to doing it as a hobby). It's a shame to see it diminished, when I believe we should be diverting efforts to automate other stuff.

Second, most AI art I've seen has been quite substandard compared to human art. We still don't know very well what human emotions are, the origin of sentience and qualia, etc.. But I think humans still lead here in having and probably understanding emotions. While for other tasks most implementation detail is irrelevant (e.g. in code, that it works tends to be most important, vs. minute choices in style), in art every detail is particularly relevant. Knowing this, it bothers me usually when I see this art that it doesn't carry the same knowledge of context and nuance a human would have.

Third, There's also the effect of making me question whether each piece of artwork was made by a human or AI, that didn't exist before. It does carry a bit of a magical feeling I think knowing a real person made every piece of artwork prior to 2018 or so (I think algorithmic art[1] is fine in this regard, because it tends to be more clearly algorithmic, and the involvement of the artist in coding is significant), that is now gone or at risk. Even the thought of imagining say their work day or what they had for lunch or talked to coworkers or friends is pleasant to me (at the risk of romanticizing it too much).

I suppose if AI art actually understood human nature, and specially the specific context of each art piece, better than us some of my arguments might be diminished. But the negatives so far seem to outweigh the positives, and I would like to e.g. give preference to content that doesn't use AI art.

(It is, admittedly, also the case that we lost a similar amount of craftsmanship when the industrial revolution happened, and in return we were able to support a larger population, and greater material conditions for most people. Every object now isn't carefully handcrafted. I think it's different because well, now material conditions are relatively abundant, and second there's no such insatiable, significant and irreplaceable demand for art as there were to common industrialized objects (take shoes for example), at least not to the same extent or vital significance. That is, the ability to have a shoe at all far outweighs it being carefully handcrafted, I believe; while experiencing a poorly made AI movie or artwork might be actually worse than none at all (or simply an older human made movie), and it also gets more cumbersome to evaluate for ourselves whether AI was employed or not. Also, while say shoes only last a limited time and need to be constantly produced, good artwork can last indefinitely (using digital storage), and even if you account for cultural change and relevance, can still last a really long time, motivating investing more into it.)

I'm quite sure that if we're still around in 500 or so years, we'll still be enjoying say Starry Night by Vincent van Gogh (probably as a digital reproduction). Current AI art will probably be largely discarded, so seems largely an unwise investment. Actually this kind of applies to code as well. It seems plausible Linux could still be used in 500 years from now (see how we still value finding Unix v4 50 years after), or at least of some interest. Those durable intellectual goods don't seem like wise places to invest anything but the best of us :) (at least in the cases it's not disposable)

The arguments above also don't seem to apply say in concept stages, or say for bland corporate diagrams that will be disposed of in 1 day, and which a huge quantity is needed. I think the main criteria I would evaluate is (1) Was it enjoyable to produce (for the artist(s))?; (2) Will it have a significant (artistic) impact on who is experiencing it?; (3) Will it last a long time?

[1] W.r.t. algorithmic art (and digital in general) (take bytebeat[2] for example), which is a field I really love, I am not any kind of absolutist about it. I know there tends to be extremely more degrees of freedom for human expression in a manual piece than in an algorithmic piece, so I see it more as a complement and not a substitute for more conventional art. I'd never give up ever hearing human musician player music for bytebeat, just bytebeat is a lovely experimental other dimension of expression. Writing a prompt seems a too few degrees of freedom and context, and too much of an uniform context that is less rich than humans can provide.

[2] https://dollchan.net/bytebeat/

Along the same lines, the anti-AI attitude among musicians today reminds me quite a bit of the anti-synthesizer attitude of the 60's and 70's, down to the same exact talking points: fears of “real” musicians being replaced by nerds pushing buttons on machines that can imitate those musicians.

I think the fears were understandable then, and are understandable now. I also think that, just as the fears around synthesizers didn't come to fruition, neither will the fears around AI come to fruition. Synthesizers didn't, and generative AI won't, replace musicians; rather, musicians did and will add these new technologies into their toolsets and use them to push music beyond what was previously understood to be possible. Synthesizers didn't catch on by just imitating other instruments, but by being understood and exploited as instruments in their own right; so will generative AI catch on not by just imitating other instruments, but by being understood and exploited as an instrument in its own right.

The core problem right now is that AI (even beyond just music) ain't being marketed as a means of augmenting one's creativity and skills, but as a means of replacing them. That'll always be misguided, both in the practical sense of producing worse outputs and in the philosophical sense of atrophying that same creativity and skills. AI doesn't have to produce slop, but it will inevitably produce slop when it's packaged and sold and marketed in a way that actively encourages slop — much like taking one of those cheap electric keyboards with built-in beats and songs and advertising it as able to replace a whole band. Yeah, it's cool that keyboards can play songs on their own and AI can generate songs on their own, but that output will always be subpar compared to what someone with even the slightest bit of creativity and skill can pull out of those exact same tools.

Electronic music history is basically a graveyard of "this isn't real music" takes that aged badly

> I'm a musician, but am also pretty amused by this anti ai wave.

Let me guess: you're an amateur musician. Not that there's anything wrong with that, but it makes it much easier to be amused about this topic.

> There was recently a post referencing aphex twin and old school idm and electronic music stuff and i can't help bein reminded how every new tech kit got always demonized until some group of artists came along and made it there own.

What are you talking about? Which "tech kit" got demonized by whom? Of course, there were always controversies around techniques like sampling or whatever, or conservatives in the UK demonizing rave culture, but otherwise, I have no idea what you're referring to.

nobody demonized afx or idm bro. autotune, yes. but that's different. damn autotune to hell

[dead]

[flagged]

[flagged]

>someday someone will come along and make a genuine artistic viable piece of work using ai

Always has been :)

https://www.youtube.com/watch?v=SpUj9zpOiP0

(And honorary mention)

https://www.youtube.com/watch?v=fYKAOPj_uts

Yeah, I agree.

For me, music (like all fine art) is about human connection. It's the artist telling me something human and personal. It's not entirely about the aesthetics of the music. The provenance of the art is very important. If I feel that connection with a song and it turns out that the song wasn't made by a person (it hasn't happened yet as far as I know), I have been deceived and would be furious.

A song made by a person using AI as tool (rather than to generate the music) is different. What matters is that the song is actually an expression of humanity, not the tools used to make it.

However, the presence of AI-generated music means that I am not really willing to buy music anymore unless it's either a few years old or I'm buying it at the merch table the artist has at a live performance.

We're in the very early stages of AI generated art. What will it be like in 10 years time? 20? 50? You might think it won't get much better. I think that's unlikely.

[flagged]

No I don't think that really qualifies because it's solving an engineering problem. I hang out on an electronic music creators' forum which is stringently anti-AI, but nobody objects to things like stem separation. People are skeptical about AI 'mastering' but don't really object for similar reasons.

What people get mad about is the use of AI to generate whole tracks. Generating rhythms, melodies, harmonies etc via AI isn't greeted warmly either, but electronic musicians generally like experimenting with things like setting up 'wrong' modulation destinations in search of interesting results. I don't think anyone seriously objects to AI-produced elements being selected and repurposed as musical raw material. But this is obviously not happening with complete track generation. It's like playing slot machines but calling yourself a business person.

That's not AI generated at all. Using acoustic models to stem out individual sections from a recording is not creating new material (and I wouldn't even describe that as "AI" despite what I'm sure a lot of the tools offering it want us to believe).

I think it makes some sense to allow leeway for intelligent "signal processing" using AI (separating out individual tracks, clean-up, etc) vs generating new content with AI.

Similarly, say, for video editors, using AI to more intelligently rotoscope (especially with alpha blending in the presence of motion blur - practically impossible to do it manually), would be a great use of AI, removing the non-creative tedium of the process.

It's not clear where the line is though. I was quite impressed with Corridor Crew's (albeit NVidia+Puget-sponsored) video [1] where they photographed dolls, motion-captured human actors moving like the dolls, and transferred the skeletal animation and facial expressions to those dolls using GenAI. Some of it required nontrivial transformative code to accommodate a skeleton to a toy's body type. There's a massive amount of tedium being removed from the creative process by GenAI without sacrificing the core human creative contribution. This feels like it should be allowed -- I think we should attempt to draw clearer lines where there are clearly efficiency gains to be had to have less "creative" uses be more socially acceptable.

[1]: https://youtu.be/DSRrSO7QhXY

That feels legit to me. We have been using software to isolate individual instruments from a recording for a while.

I think any line is necessarily going to be arbitrary, a blanket ban on any ML model being used in production would be plainly impossible -- using Ozone's EQ assistant or having a Markov chain generate your chord progressions could also count towards "in substantial part", but are equally hard to object to.

But we also live with arbitrary lines elsewhere, as with spam filters? People generally don't want ads for free Viagra, and spam filters remain the default without making "no marketing emails" a hard rule.

The problem isn't that music Transformers can't be used artfully [1] but that they allow a kind of spam which distribution services aren't really equipped to handle. In 2009, nobody would have stopped you from producing albums en masse with the generative tech of the day, Microsoft's Songsmith [2], but you would have had a hard time selling them - but hands-off distribution services like DistroKid and improved models makes music spam much more viable now than it was previously.

[1] I personally find neural synthesis models like RAVE autoencoders nifty: https://youtu.be/HC0L5ZH21kw

[2] https://en.wikipedia.org/wiki/Microsoft_Research_Songsmith as ...demoed? in https://www.youtube.com/watch?v=mg0l7f25bhU

>The source material was a compact cassette tape from a demo, found in a drawer.

Was this demo his, or someone else’s IP? If he is cleaning up or modifying his own property, not a lot of people have a problem with that.

If it is someone else’s work, then modifying with AI doesn’t change that.

I think they just don’t want AI generated works that only mash up the work of other artists, which is the default of AI generated stuff.

If the Beatles can use AI to restore a poorly recorded cassette tape of John Lennon playing the piano and singing at his dinner table, I think it's alright if other bands do it, too.

https://en.wikipedia.org/wiki/Now_and_Then_(Beatles_song)

> It's difficult to draw any lines line on a creative work, by just by nature of the work being creative.

If you want to be some neutral universal third party sure. If you're OK with taking a position, the arbitrariness actually makes it much easier. You just draw the line you want.

Creativity demands limitation, and those limitations don't have to be justified.

Suppose we have two image-oriented AI's.

One is trained with a set of pairs which match words with images. Vast numbers of images tagged with words.

The other is trained on a set of photographs of exactly the same scene from the same vantage point, but one in daylight and the other at night. Suppose all these images are copyrighted and used without permissions.

With the one AI, we can do word-to-image to generate an image. Clearly, that is a derived work of the training set of images; it's just interpolating among them based on the word assocations.

With the other AI, we can take a photograph which we took ourselves in daylight, and generate a night version of the same one. This is not clearly infringing on the training set, even though that output depends on it. We used the set without permission to have the machine extract and learn the concept of diurnal vs. nocturnal appearance of scenes, based on which it is kind of "reimagining" our daytime image as a night time one.

The question of whether AI is stealing material depends exactly on what the training pathway is; what it is that it is learning from the data. Is it learning to just crib, and interpolate, or to glean some general concept that is not protected by copyright: like separating mixed audio into tracks, changing day to night, or whatever.

That's not generative AI, just source separation which has existed way before large language models and transformer architecture were big.

You don't make it clear whether the music on that tape was "generated" "by AI", only that it was post-processed in such a way.

The example you present seems fairly straightforward to my intuition, but I think your point is fair.

A harder set of hypotheticals might arise if music production goes the direction that software engineering is heading: “agentic work”, whereby a person is very much involved in the creation of a work, but more by directing an AI agent than by orchestrating a set of non-AI tools.

This is very similar to, "am I ripping people off if I just get the LLM AI to make a few grammar fixes in my own writing?"

Ya, "AI" is too broad a term. This was already possible without "AI" as we know it today, but of course it was still the same idea back then. I get what you're saying, though: would he have bothered if he'd have to have found the right filters/plugins on his own? idunno.

thats sounds more like unsupervised learning via one of the bread and butter clustering algorithms. I guess that is technically AI but its a far cry from the transformers tech thats actually got everyone's underwear in knots.

AI unmixing or denoising is not really generating music AI.

Where does it stop. My dad is a decent guitarist but a poor singer (sadly, I'm even worse). He has written some songs, his own words, some guitar licks or chords as input material and AI turning it into a surprisingly believable finished piece. To me it's basically AI slop but he's putting in a modest amount of effort for the output.

This is exactly where the policy starts to get murky

That's not "generating" the music with AI - that's isolating the tracks of existing music. Probably not generative AI at all, and depending on who you ask, not even AI.

This is why it is to these generative AI companies' benefit that 'AI' becomes a catchall term for everything, from what enemies are programmed to do in video games to a spambot that creates and uploads slop facebook videos on the hour.

I'm a moderately frequent user of Suno and have encountered some unusual AI-generated effects that I've not come across before in more traditionally synthesized music.

One is when it attempts to generate vocals without a lyrics prompt. It's gibberish but just on the edge of comprehensibility. Sometimes it'll be entirely spoken word with no accompaniment. Very uncanny.

Another is transitioning between vocals and instrument in the same melody line. Like a humanesque voice holding a steady note at the end of a verse which seamlessly transitions into a saxophone sound and proceeds into a solo. Or vice versa, an instrumental morphing into a voice.

Finally is when the generation goes wrong and it starts spitting out absolute nonsensical sounds with no rhythm or melody, in a uniquely fragmented way I can't really describe. It feels like seeing the musical matrix, the inner thoughts of the AI.

Now I've written all that out and had a think about it, I'm tempted to sample these oddities and try to make something more structured out of them.

Yes, it's really easy to argue against AI music and find that you're making exactly the same argument that was used against DJs and sampling in the 90s. "Real musicians" thought they were being ripped off by "non-musicians" who didn't know how to play an instrument, just a turntable or an AKAI. But it turns out that turntables and samplers are instruments if you allow people to get creative with them, and now we have entire genres of music that exist because sampling is legitimate (though the copyright wars did make it more expensive than the original guerrilla days). I mean, do we seriously think DJ Shadow isn't a musician?

You could regard AI as being literally just a very advanced form of sampling. I've seen and heard some very creative uses of AI tools, and it would be a terrible shame if that baby got thrown out with the bathwater.

> new sounds and musical forms

Has it done this? Or does it just make things that sound like what it's trained on?

What I'm hoping for is a (good) musician to take a small music model, train it on a selection of their influences and existing music and then add on a prompt tracklist. Something like a NN album that the fan can add their own prompts and get new music in the style the artist meant for that album.

I don't know if this really counts as art but there are existing AI music that are at least though provoking funny like https://kommandointernet.bandcamp.com/

It is German but there is really funny remixes combining Mallorca Party music (a German music genre of its own) with anti fascist themes. They manage to make fun of both without being IMHO totally alienating both subcultures, which could count as art. IMHO such things would not really be possible without AI. Would be really disappointed if this disappeared from BC.

Most human derived goods markets need to filter low effort attempts at monetization in general. I'm tired of going to Renn. Faire / Farmer Markets / Artist Alley Markets only to find they've started letting in "Joe Blow with a poorly configured 3d printer #35." These places have become infested with people selling the exact same piles of thingverse trash in a rainbow of colors.

It sucks that a lot of these types of markets are suffering from low numbers of shoppers. They open themselves up to these plastic peddlers in desperation only to drive away customers even more.

This platform sees high front page placement / hundreds of votes for AI-written … uh sorry I mean “I used the AI to clean up my notes” … content daily.

I have an allergic reaction to it and flag, but clearly a majority of the hn voting population appreciates it, so if it isn’t banned it will continue.

No one in these threads ever discusses how you would identify and remove AI-generated music.

E.g. how is this worse and needs to be removed: https://youtu.be/L3Uyfnp-jag?si=SL4Jc4qeEXVgUpeC but crap that top pop artists vomit out into the world doesn't

I wonder why the head honchos of the music industry haven’t come together and put together an alternative to Spotify’s increasingly AI-slop platform

can't see Microslop GitHub banning AI slop

Convenient helper site: https://isitbandcampfriday.com/

I usually buy stuff on release day, but when I find older albums (or the rare ones I preorder), I wishlist them to get on BC Friday

Whenever I see defences of AI "art" people very often reduce the arguments to these analogies of using tools, but it's ineffective. Whether you use MS Paint, Photoshop, pencil, watercolor etc. That all requires skill, practice, and is this great intersection of intent and ability. It's authentic. Generating media with AI requires no skill, no intent, and very minimal labor. It is an approximation of the words you typed in and reduces you to a commissioner. You created nothing. You commissioned a work from a machine and are claiming creative authorship.

I am a musician, in the “accomplished amateur” category. For me, music is a never-ending journey of learning and skill-building, and I’ve come to appreciate that journey as much or more than the destination (= recording or live performance). If you gave me a one-click button to improve my skills, I’m not sure I would click it— I’d rather get there myself.

I’d encourage you to dig deeper into why and how the music that is being created by those tools works.

The whole problem with this is that the people using generative AI tools are trying to co-opt what "art" is. They're barging into creative spaces and demanding that the real artists treat them as equals. I hope you understand that no group of people would treat you kindly for doing that.

That’s what the sites that generate the music are for. Bandcamp is for musicians.

The community of AI "musicians" should absolutely have a place to share their "music". Just identify it as such so everyone else can avoid it. :)

An aside but https://sunoai-music.com/ seems incredibly broken on my desktop.

Typed a prompt and hit generate. No response after waiting some time. I scrolled down to existing sample music to get a sense of what it creates and hit play. Not one of the play buttons worked. Ok load up Chrome instead of Firefox, maybe they did some Chrome specific thing? Nope site's still broken and none of the samples under "Suno AI Music Gallery" actually work. There's a javascript error "invalid client" on clicking it. I'm not logged in i guess?

It did work on mobile but that seems like it presents a completely different site.

What workflow are you using with Ableton/Suno?

...and now deviantart is deader than dead because overrun with slop.

Anyway I don't think your case is really so bad. As long as the creator at least has put in the effort to listen to their own stuff from beginning to end at least once (yes that's a low bar), you're already miles ahead of people who'd auto-gen 100s of albums and slap them on there in one go. Music is more inherently rate-limiting than image generation where only half a second or less is needed to take in an image superficially.

[dead]

The music is produced by a large team and the headlining artist reduced to a brand in that case, but ultimately it is still a team of people. The use of AI in their place forecloses the possibility of their artistic drive overcoming and subverting the conformist demands of the music industry.

I’m not going to listen to people poo poohing what I’ve been doing with Suno, my seven year old and I rocked out to a song that took me less time to make than it took to listen to about my character’s D&D adventure last night. I’m just having fun.

The big secret right now is Suno will output good stuff but then it’s tortuous to get the lyrics to line up if you want to change anything, or add a word, it screws up the entire flow of the song all the way through. I spent 5 minutes making a 6 minute long power metal style song about a Druid fighting a dragon, then two hours unsuccessfully trying to get one with slightly more coherent lyrics and it output like 8 songs that sound terrible in a way I can’t quite put my finger on. The one song I tortured into existence took me 5 hours of work after getting a rough draft.

That said, being able to instantly make a song to tell my kids they need to clean the living room before we open presents and them singing the chorus happily for weeks after is just this great unique memory we have. Bespoke songs just for us is one of the coolest things ever and no amount of grumbling from anyone can dissuade me from it.

> [...]just being able to hear it for real and maybe share it with some people.

Your ability to make and share music as you like hasn't been abridged. Bandcamp has chosen not to be a part of it if it's AI-mediated.

> We want musicians to keep making music, and for fans to have confidence that the music they find on Bandcamp was created by humans.

It sounds like bandcamp is not the right place for what you want to do. There's plenty of ways to do what you're looking for though!

>I've spent many hours learning to play guitar and ukulele but I'm really not very good, and probably never will be - but I can hear the music in my head I want to create. I'm not interested in monetary gain at all, just being able to hear it for real and maybe share it with some people.

Luckily there are hundreds of MIDI editor DAWs available. Open the piano roll, and write down the music note by note. Surely, it's not that hard if you can hear it in your head.

AI is not the only way to make synthetic music. If you have an exact idea, you can use virtual instrument plugins for software like Ableton Live to produce music.

That's how I feel about it too. I play piano well enough to get my ideas out and have used synthesizers and digital audio workstation software for years. I can sit down and carefully craft the music I have in my mind if I want to. Though I don't often have the time or the mental space to do so.

Or, I can make a barebones recording of musical ideas - of melodies, harmonies and the overall song structure - then upload it to Suno and inspire its robotic session musicians to play it in the style I ask. Quicker to make, faster to iterate upon, and can even be used as the basis for resynthesizing the track using more traditional methods.

Half-finished tracks, ideas that were only in my mind or existed only as badly recorded piano-bashed drafts, now (almost) fully fleshed out. It's immensely satisfying, and has made me even more creative as I use this tool to understand and explore musical styles I'm less familiar with.

It sounds to me like you just need to try creating electronic music, not using generative AI. If you really have an exact idea of what the music should sound like, you should be able to realize your vision this way, and you can use realistic instrument samples if you want. It may not sound the way a person would have played it, since you can't fully replace a human performer, but it can still sound really good, and there's a huge body of video game music from the last 30 years or so to vouch for that.

> If I have an exact idea of what I want something to sound like, and I'm able to use an automated system to create that, is that creative expression?

i say this without malice (more as an encouragement to you to do whatever you like): nobody cares about your individual consumption choices. if you want to consume content from a machine that produces the thing you want to consume, that in and of itself has no impact on anyone but you. but you might be missing out on good music (either by someone else, or by your own hand) and ultimately, i wonder if you will value the end product of a musical vending machine.

People have been doing this for decades already, FL studio, Abelton, etc. Not sure what AI has to do with it.