Yes, it was originally a legal dispute. Universal Music Group (UMG) and other major labels sued Udio in 2024, accusing the AI music company of using copyrighted music without permission to train its models. However, the case was not decided in court but recently (at the end of October 2025) ended with an out-of-court settlement.

 

Sadly, whether these startup companies succeed or not, the cat is out of the bag. The technology itself will never go away … unless all of the models are somehow seized and destroyed (impossible). There will always be ai music from now on.

 

It is the same battle than working class has done for centuries ... and we already know how they ended.
https://grokipedia.com/page/Luddite

At the end, AI will simply eat everything it can, like industry done with manual workers.
"Virtual internet world" was the last portion of the world without "automated" revolution.
It is going to be faster than ever in human history.

1st agrarian revolution : millenaries
2nd industrial revolution : centuries
3rd services (tertiary) revolution : decades
4th AI revolution : less than a decade

Remember : all this started around 2021 only.
Less than 5 years ago.

Amazon is kicking 14 000 ppl, UPS 34 000 ... only THIS month.
Do the math in years passed and to come.
https://www.nbcnews.com/business/bu...-ai-are-companies-actually-getting-rcna240221

ppl thinking they can "stop it" are like luddites (read above).
I can already tell you "big studios" are going to loose their fight in the long run.
And actors too. And music industry too. And ...

 

The impact of replacing the currently employed with robots or software will be in the very near future much more devastating than it was in the early 1800's when looms were mechanized. There are now billions worldwide who will be directly negatively impacted. There is no plan at all of how to cloth, feed and shelter those with no income. It's just a few years away.

 

We are so done, bro. 147k followers on Insta and a multimillion-dollar record deal:
https://www.instagram.com/xania_monet/

Xania Monet is the first AI-powered artist to debut on a Billboard airplay chart, but she likely won’t be the last
https://www.cnn.com/2025/11/01/entertainment/xania-monet-billboard-ai

 

Last edited: Nov 1, 2025 at 2:27 PM

OMG sooo good

At least they fought between them for a while. I know, not much of a consolation

 

That honestly just seems like they're trying to push the narrative
The same guys and investors whose money's deep in AI music, they promised success and they still waiting for it to blow.

They simply can't profit off AI music if it isn't popular, the system always relies on the audience.
That's probably just an attempt to make people THINK they should be listening to Xania Monet, coz she has a "deal"...
Its the old trick, fake success to get success.

Timbaland is trying really HARD to push his AI artist, with very minimal success
Even with his top connections, money and every tool at his disposal to make it work.

It would be pretty insightful to know how much % of the music industry yearly revenue is generated by an audience who does not care at all about the artist... That's where AI music could tap into

Think of those AI influencers that blew up last year on Instagram. They even got paid to be in very popular brands...
(back when AI humans looked VERY fake)
since there was no such person, people never engaged about it.
It was popular simply out of curiosity, not real connection.

Meanwhile Justin Bieber can post a picture about a fly on the wall and get 3M likes. That's where the money is. numbers do matter
Of course he's a huge influencer, but the same principle can be seen at a much smaller scale.
AI is simply not engaging for brands, luxury fragrances still invest HEAVILY into getting the latest face to promote their new product.
Rihanna did 30 times her music money with Fenty, the brand only worked thanks to her influence in music

You won't hear much in the future about "Xania Monet"... it was a flop from the start
If only the song on its own mattered, we wouldn't see the same artist have 2-3 songs simultaneously on Top 50
That's the artist itself trending, not the song

Of course i could be wrong, I'm not an expert or anything in the business

 

They're working on one in conjunction with Udio.

 

Suno's next in line.

 

This is the most important point. We all know what it can do now and eventually some open source model trained on all of the copyright stuff will emerge. It can't be stopped. Maybe 100% AI generated songs can be detected but if you create in AI, dump to DAW, and your own instruments/flavor/touch it will be undetectable. These lawsuits are making tons of money for lawyers and in the long run will be ineffective.

 

If you use these ai services and then "dump to DAW", they are still detectable not only because of the audio indicators but also the midi and the file itself has identifiable data attributes all of which you cant erase without completely ruining the files themselves so using your method you must recreate everything with your own sounds and then completely delete all ai files from the project before you render to audio or it will still be detectable. and as this thread title indicates these ai companies are now working together with the streaming and sales sites licensing their detection algos or even totally integrating with the commercial platforms to flag ai songs. Another thing is governments around the world are putting restrictions on copyrights and royalties with ai content which I think will only get worse. So is it really worth all the trouble and money to "make" a song that you can never claim as yours or ever get paid anything for it? You are becoming a slave to the major label and streaming platforms who get paid for your uploaded content regardless.

At the point of doing all that to "make" a song, it is all kind of pointless I think. You might as well go ahead and "dump to DAW" your favorite song and recreate that completely from scratch then "flip" your remake to something unrecognizable, in other words a remix, but not labelled as such and 100% yours to do with what you want and never worry about any watermarks, copyright or other file ID's these ai scumbags or major labels use, who btw also own the streaming companies. What a stupid racket!


One final thought before I go "dont feed the credits monster!" -quote by, I forget who said this in another ai thread here. Youre wasting your time and money and contributing to the ongoing destruction of professional music that isnt controlled by major label system.

 

Not trying to be argumentative, but rather genuinely curious, do you have a source for the information you listed above? As far as i have found, and i very well could be wrong, when you start to chop, transpose, eq, add reverb etc it will erode the identifiable attributes in the audio. To my knowledge, there does not exist an algorithm that can withstand such destruction. If you have any further information or reading i can do please share!! As far as midi, while there may be identifiable data in the midi file itself, that data would fail to be recognized as soon as you turn the midi into audio by running it into a synth etc. Again, not trying to be argumentative. I have done a fair amount of my own research, but easily could have missed something and would love to learn more.

Im also not trying to be a shill for AI, im rather indifferent on it for the most part. If somebody is prompting AI, then releasing that direct output as their own, then ya, that's pretty dumb and i think we can all agree they are not a producer. If somebody uses AI as a launch pad to build around and then remove, or to create sample fodder to warp, bend, and chop, i really don't care. Although of course, there is the ethical factor of the AI's being trained on works not approved for training.

I actually got suno to come within 90% of accurately recreating a song for an artist i produced. The song has moderates success, 30 million streams or so, and i wanted to see if i could determine if suno trained on something i had a hand in creating. It took just a few generations of prompting and revision. I fed the same lyrics, and used certain keywords etc and yah it was too close. Even the vocal delivered by suno had near identical characteristics.

Im hoping to see more so-called "ethical AI" come about, but we'll see.

 

After spending months on Suno and spending more money than I am comfortable admitting to here, I tried asking them directly about all of this and they wouldnt answer me. They always deflect the question and change the subject using words which really have no meaning and dont answer the question about these identifiers. I paid for the product but they wont answer simple question about their products so I know I am safe when I use it? its crazy. I then spent a week searching and using ai questions + answering and found lots of links to information about the techniques they use for identifying the content they generate and sell. These ai generation companies are very secretive about this because once there is a "fix" they can no longer make money from people for credits they sell or license their software to streaming companies and stores, it becomes valueless at that point. How else can they know if your song on Youtube or Spotify uses ai parts or the whole song is ai. How else can they detect if you have free, pro or premium account and paid for the credits to make the song or parts. It isnt difficult to understand they built these models for this purpose from the start.

The conclusion is, and yes I searched, read and asked many sources about this, there is no way to remove the audio or data based identifiers from whole songs, stems or midi without completely ruining the file to the point of it being static noise. Why th pay for ai if you can only use noise or totally different sound? we can make that ourselves for free in 5 minutes Even rerecording, resampling, chopping up, transposing, using reverbs, eqs, none of it will completely erase that identifying information they embed in the files. As I said before you can only recreate everything yourself from scratch then delete everything ai generated. This is the only way to use it safely and to ensure you dont get screwed over and in trouble with the laws around the world whether its today tomorrow or 10 years from now. Its not worth the time, money or risk I think. Its better to only use ai for an idea and then do everything yourself from nothing and delete all the ai.

I wish someone told me all this before I wasted money, so I am doing the same now for others that is all!

 

But I must ask, as you give advice to "don't feed the credits monster", etc..
Why on Earth would you even try to use it, and then spin around and give moral advice?
You complain that you can't use the ai as your own creation, and then moan about how the big ones will steal the ai...
WTF.

 

If the audio leaves the computer as analog audio, and is recaptured back into the computer again via an ADC, there is no way their "special frequencies" are invulnerable to all the processing it might be subjected to by normal production work. Nevermind someone actively out to avoid detection.
Most of these people trying to pawn off generative AI output are not proactive enough to make the music themselves, so that sort of work is usually beyond their level of effort.

 

Soon we'll see more and more tools that eat up part of the production process.
Whether that's an IN-DAW MIDI writer or an AI that can generate drum samples.

Most likely the developer of such tools will want a percentage of the royalties, or charge $100-500 upfront for the app. (like current AI vocal plugins)
Otherwise why bother making a tool you can't profit from.
open source exists, but due to the complexity of music, you have to be a get a HUGE amount of data for it to work.
Even if you get 1 billion songs, you then have to process those 1 billion songs (big hardware and power costs)

I'd be much more concerned about OpenAI, they're working on their own music tool and they've proven to make good stuff.
And that's because they have the resources.
Sora was a hard hit for VFX artists and cinematographers.

As long as these tools don't go absolutely superior to every human producer, they'll remain as something like Scaler 3...
It can make good chord progressions, but someone with basic music theory will make something far better.
Chords are one part of the song, if you need a tool for that, you probably can't build the rest professionally.
There's always exceptions, but a pro musician has a higher rate of quality output.

So yeah, AI might make you a drum loop or give you a guitar player to play along. (Logic already does something like this)
But it will most likely just slow down the production process for many,
Its very common to second guess the things you didn't create, especially if you can hit "create" again, and again, and again...

Think about Splice, that's ALREADY an infinite music machine.
But a loop-based song vs fully written song has no comparison in quality.
Sabrina Carpenter "Espresso" was famous for using a guitar Splice Loop, but not because you use a splice loop you'll have a radio-ready song.
Most Splice users can't exit the 8 bar loop.

Professional songs have EVERY single element on it, with purpose and space.
And that is super important to keep in mind, purpose and space lives within harmony, color, time, mixing and few other abstract stuff.
I personally think only a human ear can tell what works and what doesn't. (when, in combination with what, etc)

Plus one thing is the production and instruments of your song, but then mixing is also heavily creative.
Just FX and automations already require a huge time and effort investment to get it perfect.
You're not looking for a clean sound, you're looking for YOUR sound, for that song.

So back to my first point, most serious artists and labels will keep sticking to CLEAN songs.
No one wants to use something that already has 50%+ taken by a third party.
Unless that 50% belongs to themselves, like UMG.
But if that's the case, producers won't use the UMG tool because it cuts their paycheck.
Artists will stay working with people that have a trajectory, like Serban Ghenea, Rick Rubin, etc... they keep appearing in big projects over and over again because their work and success speaks for themselves.

But its just my opinion, all pure theory and speculation

 

wow you seem very offended by me telling the truth! Allow me to simply for you. I learned from my mistakes and share my experiences with others so they dont make the same mistake. It is called sharing wisdom. It is what forums are designed for. Understand?

i do make my own music, and I play several instruments proficiently and professionally, but Suno was sold to me as a sort of "super daw" and way to spark inspiration and also help create parts for my songs using my own creations along with Suno Studio. It turns out.. well, you can read my previous comment to learn that


If you want to discuss morals, Suno has proven they have none. If you are defending any of this, then you are part of the problem, maybe even work for them or get some affiliate perks? Anyway i dont know why else you get so offended by the truth, calm down buddy. Stress is not good for your health!

And if you dont like reading the truth, then maybe go to another thread?

 

Sure you can place your phone mic in front of your speaker and record the audio then import back into new project. But there are still audio identifiers in the file. Like I said, there are ways to erase these but not without completely destroying the files. This is not only my own tests but facts presented by way too many sources to quote and link here (and in another language to boot!). My whole point is why waste time and money on something that you cant use and must completely recreate yourself anyway. You are making music creation life very hard and expensive on yourself and will not be able to benefit from or recoup anything if using ai generated content in your published songs. These are not my facts but facts all the same guys. But of course you can take it or leave it and ignore to find out for yourself!

 

this might not be the best example, but you can run a song effects, processing, transpose... and content id eventually strikes.

Even so, why bother so much trying to bypass the system by altering the product?
You end up with trash anyways. I prefer to listen bad stock music in social media, than a high passed, pitched up version of a popular song in the background

And we saw this happen with youtube
In earlier years, people found ways to have copyrighted content on their videos and still monetize
By flipping the image, changing colors, adding zoom ins, etc.
At some point they get caught and monetization gets removed, not worth the risk.

People find ways to bypass and they just patch the hole. Like the Adblock fight