Это видео недоступно.
Сожалеем об этом.
Is Suno Sampling Copyrighted Producer Tags? (Audio Examples)
HTML-код
- Опубликовано: 17 авг 2024
- #aimusic #suno #sunomusic #djkhaled #jasonderulo #nickmira #cashmoneyap #copyright #mikewillmadeit #djmustard #lawsuit #riaa
0:00 Intro & Lawsuits
1:15 Suno Getting Sued
2:55 "CashMoneyAP" Tag
4:57 Producer Tag Examples
6:28 Compilation Tags
7:44 "Daytrip" Tag
9:13 3 Questions
10:05 Sampling on Spectrum?
11:21 Value of Data in AI?
13:46 We hold the gold
My Udio video: • Is Udio Reproducing Co...
Wanna learn more about TV/Film Sync Licensing?
www.syncmymusi...
lmao yea this is the same as when they caught midjourney replicating artist signatures
@@Essenger or Dall-E adding stock photo vendors watermarks 😆
the difference is who you're messing with - the music companies have money and armies of lawyers. 😏
You think the record companies etc want to be able to use the music they own, not udio or anyone else, to generate their own AI ‘assets’ to plunder the works of songwriters artists and producers. Historically they’ve never been altruistic towards the people who create the music.
Like most things in life... It comes down to... It is easier to ask for forgiveness than permission.
Suno et al know exactly what they're doing! Sadly. 😒
This play might bankrupt the company, basically it's a legal smoking gun. But yeah Suno knew the risks and I'm sure are just trying to ride out the lawsuit without going under.
Just unbelievable….i’m still digesting this…it’s not even subtle. They don’t even seem to be synthesised versions. The tags sound exactly the same. 😳 how exactly does Suno work again?
it is obvious that these companies are breaking the law, but it is also hard to imagine that these sites and technology will be destroyed just like that. perhaps the parties can come to some kind of compromise, as people are willing to pay money for this service, which means it is an additional source of income for copyright holders and labels. but let's hope that the laws are strict, because this abyss of music will be an unhealthy competition.
It is not obvious. Record companies will either buy these companies or create their own similar AI music generators
I just tested Udio by prompting it specifically with "cash money AP" as an intro to a rap clip - it didn't take the bait, but... on 1 out of the 10 generations I requested, interestingly, I received a "moderation error". 🤔
Now what does this tell us? Well, I ran the same prompt with the same settings several times over (Udio generates 2 clips at a time) and only one of those were rejected with a "moderation error" - and also, this generation took the same amount of time to finish as the other ones. This tells you it's not the *prompt* being rejected, it's something else - something happening *after* the output has been created. 🤨
The likely explanation: they're now checking the model's output and testing it for similarity with well known vectors, which derive from clips of copyrighted music that they don't want the model to recreate, e.g. when they receive complaints or bad press for cloning copyrighted music.
There are so many interesting things you can conclude from there. First of all (not that there was ever any doubt) that means the model is capable of recreating fragments of copyright recordings - which means (again, not that anyone doubted this) that it was trained on copyright music.
If you understand who AI works, you already know that training a model like this is very, very expensive and time consuming - millions of dollars and months of computation. You might also know that there's no way to remove something from the model - the model is "holographic", in the sense that "everything is everywhere", you can't really separate sounds or melodies from each other, or from anything else, it's all "one thing". (technically, yes, there are ways to do it, but it "damages" the model, lowers the output quality - you can "erase" something, but not without damaging something else.)
Since they can't remove anything from the model (and since they can't retrain the entire model without the recordings they've been caught stealing) the best they can do is (A) try to prevent you from asking for things they shouldn't generate, their first line of defense, which they've been doing for a while, and (B) try to catch themselves when they do create something they shouldn't have, which they appear to be doing now.
The problem of course is these counter measures are not designed to undo what they did in the first place - they're only designed to hide the fact that they did it. To plug the holes, so to speak. Or to bury the bodies, if you will. 😌
@@RasmusSchultzthanks for sharing. I was thinking the same that now at this point since they've been exposed the only thing that can be done technically in the output validation stage is write conditional checks against the same database storing the audio and meta data used to train the models in the first place.
if (sounds like Whitney Houston) {
return "sorry can't expose a crime"
}
If they were ethical and smarter they would have prioritized implementing sound generation copyright protection safe guards from the start but greed and deception was more important evidently.
@RasmusSchultz so in other words it's like asking Bing or Chatpt to create a violent image. The system will refuse to accommodate the request. This is nothing knew. AI models have had restrictions on them pretty much out of the gate. Otherwise you'd have some pretty nefarious images floating around the internet. AI music is no different. The biggest thing here is that you couldn't reproduce this videos claim. Why? Because the results were "coaxed" using certain commands and tricks. Basically BS.
@@mephistro no, if you want a comparison, it's not like asking for a violent image - it's more like asking for an image of Iron Man and getting something extremely close to the actual Iron Man movie poster. The model contains a compressed representation of that exact image. There is an ongoing case over that as well.
@@justin.johnson yeah, except I don't know if they could have. there was a paper released this week that demonstrates a new technique to reduce over fitting, so it's an area of ongoing research still. even then, there's no saying they would want to - I mean, the model has to be fine tuned to favor patterns in the most popular music, otherwise it would just produce so-so music, people wouldn't like most of it. same as how OpenAI is/was being sued for fine-tuning on NY Times articles - producing output that more closely resembles the subjectively highest quality input is not by accident, it's very much by design.
Completely by design. To "achieve" "the best" model in competition it would have to simply plagiarize what it's being marketed to do, be artificial lol
Makes you wonder how many other samples Suno reproduced that are less noticeable/identifiable. How many uncleared Splice samples did Suno spit out already? Without full licenses for everything that goes into it, it’s a minefield.
💯
kind of disgusting how these AI operate, its a similar problem with how every day of our life is archived through our data and sold off to companies like these to do who knows what with it
the big 3 sue to get a piece of the pie.. not to kill the technology and protect the copyright.. will end same way like with spotify.. artists will get scraps from labels instead of shares in these ai companies
Maybe I haven’t got enough ‘skin in the game’ in terms of copyrighted music but respectfully I don’t really care about this kind of thing. I use suno as tool to help me write. I would never directly release anything I generate with suno and I think anyone that’s even slightly serious about music would either
People also don't realize there was a ted talk explaining that melodies naturally will overlap, and copyrighting for so many years actually will stiffle original ideas, especially as most commercial songs lock towards the E key, especially pop+[genre]. This is where science/math conflicts with politics/laws, and why someone has already made all melodies possible to make public domain licensing. So really, copyright of melodies/patterns died years ago. Then we have companies who probably knew this still want to file lawsuits. Of course, producer tags is an issue, but it's no different from stable diffusion early models showing trademarks and later on evolved from it. Most AI models have users test and evolve, including GPT. But because this is under fair use at a small sample size, unlike Udio's Mariah Carrey which is more lengthy, you have no issue here. The issue should be if someone tried to actually profit off of making sales from their generation using producer tags, which none will. However, as a generator, they shouldn't be sued for what other users choose to do. That's like saying Google should be sued.
Exact versions of a song should be copywritable, not melodies. Beats already lost this battle in IP, so should patterns. Patterns forces anyone with a random idea to be hit with a lawsuit, when they probably never heard the other song with different instruments and genre.
Copyrighting beats wouldn't just be possible. They follow so trivial rules that you would crash into legal issues constantly. And indeed RUclipss Content AI claims it all the time. Even with stuff you do entirely from scratch without sample based instruments. It is super stupid.
With your logic, copyright should just be done away with.
@@mater5930 That short amount of thinking over what i Just said is foolish. Did I tell you that melodies are already in public domain?
What should be copyright is the actual exact song, not the pattern. Anyone can replicate an E-note formula if it's the only thing that sells above platinum.
Also, record labels want to already create their own AI service, which means suno is pushed aside for an even worse product, with now their touch of ownership to already public domain patterns. This gives then more power than you ever seen before over indie artst, while still owning spotify shares to have board members.
Deductively, they would literally bring back the 360 deal for pennies on the dollar more than it already is, creating debt slaves for those who want to make it in music to a livable wage even. Their mission is now about dillution/quantitiy to streams for it wins for them, not the artist. Notice music videos are less invested, so quality in the past might become a thing of the past, down to even A&R as industry profits still shrink but enough to sustain them...
So... when you put it all together, what use would your hypoerbolic question be if there's a bigger problem at hand to argue over the logic and reasoning behind your assumption that copyright would be done away with beyond the music industry but as a whole?
@@samthesomniator I agree. It overlaps too many times. Melodies will naturally find a similar. AI would probably find songs sounding similar and end up in many litigious battles. Some have already in fact and never even knew each other's existences. I understand lyrics, actual song exactness, but melodies, like instrument/tool choices, should NOT be copyrightable due to the nature of music's limitations. So long as they have lobbyists to even stiffle the existence of a music union, tells a lot.
@@mater5930 shoukd NOT sorry
You pay a museum, to be able to go in and see the paintings. Some are free entry. The museum pays the artists for the exhibition, to draw people in. I go in, I see 5 paintings, I create a mishmash exhibition of my own. Not a problem. I buy music. I listen to 5 songs. I create a mishmash of those 5 songs. I record and put out my song. Not a problem. The problem comes when I sneak into the museum and don't pay. Or I pirate the songs. If Suno paid for downloaded copies of every song, then the artist gets money (granted it's a disgustingly small amount and completely unfair but that's another discussion). If they pirate it, then there's a problem.
What's the law say about the music they acquired? Did they use a permanent copy? If they streamed it, were they allowed to record a permanent copy? If not and Suno did create a permanent copy, that's a broken law. It's pretty black and white. If the act of missmashing becomes a copyright offence, then how many human artists will be sued?
Copyright doesn't protect an idea. It protects a tangible medium of expression of an idea. The actual object on a drive or tape or paper. Until if becomes tangible, there is no protection. So if they recorded a couple of words themselves, emulating a beat or 2 of a song, and did that for every single bit of music, and then wrote out songs as notation and trained the data on like notation and written music to be able to play, that's not a problem. Any human is allowed to do that. The AI could use the written analysis to compose a song and then use the samples they recorded to piece together a song. I personally would feel like I was doing something wrong if I sampled other people's songs, split them up into bits, and then created a song from the mini samples.
@@sirhammon your brain is not an AI. Humans aren’t robots. Our minds don’t function the same way or produce the same outcomes as generative AI. And the two don’t carry the same legal protections.
This AI is doing a spectral analysis and copying the same frequencies, as this demo shows. It’s not learning a new process to create a similar work. It’s plagiarizing the “tangible medium” of the recording.
A bunch of students 10 years ago generated every possible melody and copyrighted all of them. Its like musical library of babel. So technically they can sue anyone
@@andrewlloydpeterson won't hold up, as they generated them with automation, rather than with direct human creation. Copyright exists for *human* expression.
@@NathanaelRouillard what do you mean it wont hold up? They copyrighted it. Its done. They patented it. How can you prove something is created by a human or not? Music algorithmic generators existed since when mozart was alive before even computers existed
How come you're the only person coming across all of these? It's starting to be clear the so callede "user" is more than likely you putting in prompts to come up with these because I've been using Suno since it came out and not once have I gotten any copyrighted material. There's clearly a chip on your shoulder.
@@tinagomez4418 swing and a miss. Wanna try again?
@@SyncMyMusic No thanks! with that response, I'm pretty sure I hit the target.
@@tinagomez4418 nobody believes you believe that
@@tinagomez4418 No, you didn't. Only in your head.
He has a community and people are sharing with him which he says in the video.
Calling this stuff artificial intelligence was a mistake. It's not intelligence. It's algorithmic recreation. And it's a giant sink and waste of electricity.
precisely
I agree
Agreed. With such bogus terms its like The E3 hype game trailer is a metaphor for Silicon Valley right now.
It's AI, a brain simulation.
@@abram730 it's absolutely not even close
The ML models are basically holograms. So depending on how you slice it it gives different views.
I love the sampler analogy… especially around the right side of the sampler spectrum. This is exactly what beat makers do. Chop, slice, alter and vary to create a “new” sound from a modified pre-existing work. I think Suno/Udio are great tools to create sample chopping material… depending on the outcome of the lawsuit. :)
OBVIOUSLY Suno and Udio are training on copyrighted data. That’s the crux of the lawsuit. :) The fact that people are able to draw this level of detail out of the tools is quite interesting. I think it’s worth noting that every one of those producer tag examples included the producer tag in the lyrics and/or the title of the prompt.
I think another analogy for these tools is a musical search engine. They dig into their training data to return new wave forms based on their training data. The tools are going into their training data and delivering what the prompter is asking for. This all gets back to the point I keep making about end-user actions and responsibility in this whole kerfuffle. The AI tools are not, to our knowledge, keeping an archive of copyrighted works... that music is only used during the initial training process. Prompts generate the creation of "new" music based on the training results, not the original content. For example. re: the tag stutter… the AI training apparently learned that the stutter effect was part of the trap style, or is included as a learned trait in Trap.
The plot continues to thicken!
@@YoPaulieMusic “to our knowledge” is def the issue at hand. If they’re not keeping an archive of the music/recordings, how is their model able to spit out what sounds like replicas of these producer tags? I def need to do more digging to understand these models better, but the AI companies claims that “no data is stored” doesn’t seem to ring as true considering what we showed in this video.
And no “thank you” for revealing the prompts this time? 😎
It's really cool meta did get their licensing in order. Between both suno and Udio I have 100 unreleased tracks. I wonder how many users haven't released because of the copyright mess? 😮
It could not be more obvious!!! Unbelievable! As always, great work!
Again, to me it seems like:
1. User inputs copyrighted content, infringing on holder's rights and violating terms of service.
2. User "direct the Services to generate any Output that (i) infringes any intellectual property or other proprietary rights of any party", again violating terms of service.
3. User then suggests that the service (Suno) is responsible for the resulting infringement of copyright?
Am I missing something?
EXACTLY 💯 💯 💯
Suno has a pop up warning NOT to do that.
This video is dishonest, he even covers up the propmts in the BOTTOM RIGHT CORNER of the screen so you can't see that's exactly what's happening.
WHO PUTS AN INFO BOX IN THE BOTTOM RIGHT CORNER OF A VIDEO??
😂
Like nobody would notice that.
Spot on comment.
@@BerkmanHouse yes. User isn't directing the service to infringe intellectual property, but coaxing the service to reveal the intellectual property infringement.
At 3:33 it shows that the musical prompt was just "mellow trap", no? So I'd say Suno is infringing on rights, not Suno's users. Those terms of services seem to be just meant to cover their own ###. Also, regardless of what a user inputs, this type of AI tool should never be able to spit out exact producer tags (or melodies/lyrics/beats/voices, etc for that matter) as that clearly proves it was secretly trained on (huge) amounts of copyrighted music and this infringes on holders' rights on a massive scale. It would not know or be able to replicate/imitate these tags if it was not trained on cr music.
@@JeroenvanOlffen Yes, in the first example it would seem so. But in every other example, the user seems to be entering the text that produces the tag. Terms of service are to cover their ass, yes, to protect them from misuse of their service. "spit out exact producer tags" Is it doing that? "trained on (huge) amounts of copyrighted music and this infringes on holders' rights" The fact these models were trained on copyrighted music isn't really in question. Whether that training infringes those copyrights is up for debate (and litigation).
@@NathanaelRouillard I disagree. Giving it the text and telling it to produce the tag seems like directing rather than coaxing. What intellectual property infringement by the service do you think this action by the user is revealing?
You are doing the lords work here Jesse, keep it up
Nah
So, to a guy like me that doesn't know music, Suno is a lifesaver, however, to avoid the risk, if I do a cover, I call it a cover and make sure I don't violate copywrite.
Though you cannot e.g. use it in video without permission of the copyright owner AND it has to be a cover indeed, not an arrangement.
How is it a lifesaver for you, when you "don't know music"? Then learn it, or do something else you're actually good at. I really wonder why people actually want to avoid the work to create something. Not to say with ai you're not creating anything really. The ai does. It's like telling a street musician to play a song about rainy weather and then being proud of yourself when he does. It's absolutely ridiculous. There is absolutely no pride in that. Struggling and then overcoming that struggle is the best part of the whole creative process, because that feeling is extremely satisfying. Becoming good at something is hard and that's why we give art value, because artists actually put in hours of their lifetime to become good at something. Sorry mate, but all i see in you is a joke. People like you are the worst.
You're a joke mate
I don't want to see suno get destroyed, but you make some good points. To make sure artists get paid, and so long as I had a license to distribute commercially any generations without fear of being sued myself, I wouldn't mind paying a bit more.
Thanx to people like you and Jay things get rolling in this matter, keep up the great work!!!
so what. google, microsoft etc are doing this, but with text AI. do you think they asked all the websites for permission? you don't think that's copyrighted material they're using for training? Neither MS or Google are exactly some small and shady underground business :P
Sorry to break it to you but the cat is out of the box, you can't stop this :) and as a musician i could not be happier, coolest tech since sampling for sure, maybe even since synthesizers... in a year or two it will be even more amazing than it already is. udio whatever, doesn't matter. if for some weird reason they're treated differently than Microsoft then this will go underground, it will always be a thing from here on. The future will be pretty amazing indeed.
Interestingly Google did pay for the music it trained it's AI music model with. Why do you think they would waste money on something not necessary?
@@SyncMyMusic they'll have to answer to that, there could be a numerous reasons why they selected a different route with music in comparison to text. Maybe they wanted music that sounds in a specific way.
@@Vurt72 Meta also paid. So did Stable Audio. Sam Altman has stated that OpenAI won’t be doing music because of the legal issues. Draw your own conclusions.
@@SyncMyMusic again, i really don't care at all what happens to udio and whoever else making money of this. The tools they're using is what's interesting to me, and they're not going anywhere, they will only get better with time, like with all other AI. what i really want to do is train with my own audio / my own selection of it.
Stable Audio sounds kind of garbage though, you can directly tell they have a very tiny library. But they'll get better with time i'm sure.
@@Vurt72 you mean you want to use these kinds of AI models consensually with your music? Cool, then we’re on the same team. 👍
AI music generators are just super-samplers and have the same legal issues.
@@scottbuffington5964 well it kinda does. AI certainly does not sit in a studio recreating, mixing and mastering every sound from scratch.
Edit: I mean... did you even watch the first 25 seconds of this video 😂
@@fftunes I could say that humans are just sampling software, and ask If you ever watched police academy.
I've been using it for a while, I haven't had that issue but then again I'm using my own lyrics sooooo... :)
The labels have more money than these companies do. If they really wanted to pull a boss move they would just purchase both companies
They will either purchase them or take them down and create new ones similar to them
Careful, Jesse... Anyone can upload music to Suno now. If someone uploads a song with a tag, that's probably how it really got there. Also, when you upload something, it stays in the database and you've given the rights for others to use it. That's in the terms:
"By using the Service or OTHERWISE TRANSMITTING SUBMISSIONS TO US, you grant to Suno and our affiliates, successors, assigns, and designees a worldwide, non-exclusive, fully paid-up, sublicensable (directly and indirectly through multiple tiers), assignable, royalty-free, perpetual, irrevocable right and license to use, reproduce, store, modify, distribute, create derivative works based on, perform, display, communicate, transmit and otherwise make available any and all Content .... "
In short, these tags may not be coming from Suno. All of those tags could have got there from someone adding it to Suno. Then, if you intentional prompt for them, you're going to get them at some point. That can happen accidentally to. there might be a number of people ignoring the warning about having the ownership of any uploaded material. Suno SHOULD be screening for that, but may not be, which is problematic.
I'd be careful making this claim, or any others, during a lawsuit unless you want to be sitting in a courtroom testifying to something that doesn't turn out to be true. Just trying to keep you safe. Both platforms will have to reveal their data sources in court during the discovery phase. I'd relax on this. Tha answer will come soon enough.
@@alexadigitalradio what claim did I make?
@@SyncMyMusic - The claim that the tags were created by Suno from copyrighted music they used for training is what I got out of the video. Perhaps that's not what was meant, but it can be interpreted that way for sure.
@@alexadigitalradio watch it again. I’m careful not to make definitive claims. This isn’t my first rodeo.
@@SyncMyMusic OK. Again, just trying to help as my day job does deal with countroom situations and evidence in law enforcement. A claim that isn't definitive is where it can become a problem if someone, perhaps not you, takes it to court. I'll bow out from here. Good luck down the road.
lol you really think DJ Khaled is uploading his tracks to Suno so they can plagiarize without attribution? If not, the machine is then copying without permission. Doesn’t make it legal just cuz some rando uploaded a superstar’s top single
Why don't you share the link to these Producer tag Suno files so we can hear directly from it?
@@mecd4167 email me at jesse@syncmymusic.com and I can make them available to you.
Who gives a FK?!?! Unless ur trying to release someone’s songs or material I don’t see a prob.
Ur just enabling the big 3 greedy corporate monsters to further enrich themselves
@@SinfidelityMusic I’ve got a career making copyright protected music so I give a f@ck. How about you?
@@SyncMyMusic and did someone copy and release ur music? I doubt it
@@SyncMyMusic I can also DL and sample the shit outta anything.
Doesn’t mean shit if i don’t release it or try make money from it.
@@SinfidelityMusic Thanks for your confession.
@@SinfidelityMusic my lawyers now know were to find you 🤠
Answers to your querstions:
1) There are beautiful studies shwoing how LLM's neural networks are suspiciously similar in efficiency and few other aspects to a lossy compression algorithms. So yes, technically they don't "store" any original data, but they store their kinda compressed fingerprints embedded inside the network's model's weights. Essentially they store a recipe to make something very very very very similar to the original data.
2) Yes. They did. They can't deny that forever. That makes judge's descision VERY dificult, tho. Because if they decide that a license should have been present, you'll see essentially a copyright shitstorm from every angle. Very few AI companies will survive that. The whole bubble might burst.
3) Yes. I call it a "remix machine". And if judge decides that copyright isn't needed for it's operation, I'm personally stopping to care about any copyright when remixing anything. If machine is allowed to do it, I want to be allowed to do it as well. And I won't be hesitant to use that precedent if they sue me.
💯
You should not misinterprete this lawsuit as in favor of artists. 😅 This is the music companies do to eliminate competitors and after that train their own legal AI music generators based on their database.
And as you already know. This companies are not helpful for small artists at all.
What's the better course of action that you would have liked to see?
@@SyncMyMusic I am thinking about that for a while. I wish I had an easy one. But as far as I see, and not only music and creative jobs are effected a wider rethinking of economic and societal practices will be necessary.
As it comes to creative people I think a lot of sources of income will shift from simply producing something to the actual performance and celebrity. Musicians will mainly rely on a strong (para)social connection to fans. And models like Patreon will maybe be the only way to make a living. I know ist not everybody that much an extrovert but for the moment I see that building a vital human connection is the only thing a computer simply can not provide and the best thing artists can do. 🤔
@@SyncMyMusic I also have a double sided view ob the entire topic. I am working in content creation (video editing) as well as aside from that in ambient and background stock music production. Last one will likely be obliterated by AI music generation.
At the same time I feel myself how incredible practical those tools can be in video production. Because no matter how fast and good I am at doing the score myself I always stick to my certain style and genre. And searching for fitting stock music it also tricky in comparison. Often I add some stuff like a second voice of the tracks or recollage certain Elements. AI also transformes tunes into MIDI so you can switch instrumentation. The creative possibilities are simply endless when used as such a tool in production pipeline.
If a certain band would provide an AI iteration producing convincing consistent music in their distinctive style. Alongside with their current state of live performances. I could imagine people buy this software as well.
There are open source versions of Udio and Suno. Taking them both to court isn’t going to stop this AI tech. 😂
This video is jaw-dropping. You're doing great work Jesse. These companies need to be held accountable, this is disgusting and disturbing.
Those stupid producer tags might end up saving our asses
Now tell me...Did you use the suno audio upload option to make this material on YT... To make us all think Suno is sampler?
No, just prompted these with text.
Suno knew this was going to happen. Su - sue - no = SUE NO 👀 Stay woke, folks 😂
unreal! what a video. Get it Jesse!
Artist are cheering about this, yet don't realize they had better options with suno than a label on this (who said they're making their version but for licensing). They practically can copyright every melody pattern easily in 10 years if suno and udio are removed. Even though a researcher has made every melody possible to make in under 2 TB public and free to license.
i accidentally found an what sounds like an producer tag in one of songs that suno generated
Feel free to email me it if you'd like jesse@syncmymusic.com
"Producers" that use samples complaining about being sampled.
🙃
@@HolocoughSurvivor producers who pay for the licences to use those samples, complaining about being sampled without compensation.
Fixed it for you
it all sounds like fair-use to me.
Bro, you just typed in the same lyrics. You probably went through 100 tries in order to get something close enough to sound similar. I don't believe that this was accidental. You can type anything in ( ) and it will create a producer tag. Then, use one of many apps to alter pitch and inflection. This is 100% BS and will be proven as such in court.
I don't need you to believe me, but these were very easy to prompt. For the ones I personally promptef, they came out within the first try or 2.
@SyncMyMusic you're right, I don't believe you. I've been using Suno for months and know a lot of the little tricks to "coax" the algorithm into doing certain things. It will not create the producers label unless you type in the prompt to do so. But hey, why should anyone listen to some rando on RUclips? Go try it yourself since it's free. See if you can reproduce this videos testimony.
"Bro, you just typed in the same lyrics."
Yes, but he did not dictate the speech patterns (tone, pitch, cadence, etc.). He did not prompt the model to say the producer tag in exactly the same way the real producers do. The speech patterns had to come from somewhere.
@heypfist like I stated earlier, pitch, tone, inflection can be changed in other apps. Heck, right on YouCut, the app that most RUclips videos are edited with, you can do all of that.
@pfist13 I've made almost 2000 songs using Sonu. You can DEFINITELY manipulate the inflection speech patterns in the vocals.
The Major Labels are suing to take the still the ai technology
I'm willing to bet those producers might also have a valid lawsuit as well...
We appreciate you so much Jesse thank you
If I wasn't a lifelong musician/songwriter/producer AND a decades long Senior Full-stack Software Engineer engineer that I am, NOW caught in a pickle, I would have a long time ago thought about ripping all my hundreds of CDs (Compact Disks) into mp3s, all the thousands of mp3s downloaded from Napster/Limewire/Torrents many years ago, scraping the Internet now (let Spotify run on auto play for 30 days while recording it to your DAW), and any other source and then use TensorFlow and other Machine Learning tools written in Python to train into music models and then start a for profit tech company in 2024.
That said... As a senior software engineer AND veteran musician songwriter producer they absolutely have AWS S3 buckets in multiple global regions and relational databases full of raw audio file data, meta data and documents running on automated pipelines on auto scaling highly CPU and server resource intensive machines.
Duh!!!!
It's not that easy to develop a generative music model. It would take years of specialization to learn how to do something on the level of Suno/Udio.
Go, Jesse! I think what you are doing here is great and very helpful for us composers. And your evidences are crystal clear.
Suno, etc, are OBVIOUSLY trained on copyrighted material. Without permission or payment.
Humans learn how to play guitar and learn songs they love from copyrighted music. Then they learn their own songs. You could argue that almost every song would have similarities to at least one other song. It’s no different than what AI is doing. The AI is still in progress and will learn to never reproduce exact samples or sections of the songs it’s trained on.
@@domehouse79 Writing original music does NOT distill down to just copying and pasting parts of other songwriter's songs into a new song and calling it yours. You said, "will learn to never reproduce exact samples..." ,errrm, where did you get that idea from? It's simply your opinion with no evidence.
do you think chatgpt and the numerous other text-AI's are trained on material with permission. They're not. Google, Microsoft, both gigantic companies are doing this.
Just like humans duh
@@thaddeuscoreauhhh have you heard of Sampling? Its a thing that really exists where you take other people songs and transform them into new ones. No one calls that stealing
The cat is out of the bag. If these sites are shut down or limited on how they generate music, tech nerds will release this technology for free with no limits. It's like when they shut down Napster and hundreds of other file sharing sites popped up.
And how many people still use file sharing sites instead of just streaming/paying for their music now?
@@SyncMyMusic streaming sites are more convenient for people who don't won't to take the time to download mp3s and put it on a thumb drive. A Lot of people don't Even have a computer and only use their phones. Music consumption is also dying. Country music is the only music with its head barely above water right now. Background music will be AI generated and as the technologies get better the record labels will use it to put out new hits from old artists. Welcome to the New World Order.
You are just nitpicking. The real point is:
Computers can make high-quality music now (and they get better every day).
It doesn't matter how they make it or what they use.
So why doesn't Suno or Udio just tell everyone where they got their music? Why hide that?
@@SyncMyMusic Because that's how life is. Not everyone speaks the truth, and not everyone will have the same ethics and morals as you.
And what is the point in knowing that? Will it change the fact that they know how to make music faster and better than us? No.
Get over it.
@@akazicprod the RIAA and major labels seem to think it would make a difference. Let’s see how it all plays out.
@@SyncMyMusic I can't sell you a car and then dictate to you where you can and can't drive it. If you get a driving job, I can't get part of your income because you used 'my' car to learn to drive.
6:48 If these producer tags are copyright protected, is it infringement to type them into Suno as inputs without permission? Doing so, seems to directly violate Suno's terms of service agreed to by the user. (Similar to typing the lyrics of All I Want for Christmas into Udio without permission of the rights holder). Doesn't US law put the blame on the user, for using a service to violate copyright?
Suno: "It is very important that you only upload, post, publish, or display (hereinafter, "upload(ing)") Submissions that you have rights to use and provide hereunder. By uploading any Submission, you represent and warrant that: you have, or have obtained, all rights, licenses, consents, permissions, power and/or authority necessary to submit and use (and allow us to use) such Submission in connection with the Service, including for the purpose of generating your Output."
"Submissions or uses that are illegal or prohibited by Suno. Suno reserves the right to investigate and take appropriate legal action against anyone who, in Suno's sole discretion, violates this provision, including removing the offending content from the Service, suspending or terminating the account of such violators, and reporting the violator to law enforcement authorities. You agree to not use the Service to:
submit, upload, transmit or otherwise make available any Submissions or direct the Services to generate any Output that (i) infringes any intellectual property or other proprietary rights of any party; (ii) you do not have a right to upload and use under any law or under contractual or fiduciary relationships;"
@@BerkmanHouse do you think the owners of the producer tags I highlighted in this video are going to take legal action against me or my source? Cause they’d be the only ones who could. I’d take my chances that they won’t be too mad at me.
Lolll weakest defense ever. “Don’t input the title of a real song in my song generating machine and it won’t violate copyright laws.” The whole point is that this reveals they are training on copyrighted material
@@SyncMyMusic I can only guess, but I'd guess no. I think your use of their property is fair and they don't appear damaged by it. It seems fair use if you are illustrating that the service was trained on copyrighted material (which I didn't think was really in question).
My questions concern the nature of copyright infringement in these examples. If there's infringement in the output, but it's a result of infringement in the input (by the user); then surely the user is at fault and it's not fair to blame the service (Udio/Suno)?
From Suno's view, the service is allowing use by the user, because the user agreed to their terms of use of that service. If the user then violates their agreement and uses the service in exactly the way they agreed not to do, it would seem reasonable for Suno to "be mad" at the user. (Especially if the felt their reputation/brand was being damaged as a result)
To me it seems like:
1. User inputs copyrighted content, infringing on holder's rights and violating terms of service.
2. User "direct the Services to generate any Output that (i) infringes any intellectual property or other proprietary rights of any party", again violating terms of service.
3. User then suggests that the service (Suno) is responsible for the resulting infringement of copyright?
If so, that doesn't seem a fair assessment (or fair use of the service).
@@BerkmanHouse they are liable if they don't remove that copyrighted data. If they remove it when made aware of the infringing data, then they are safe, but it seems they can't really remove it, so it is apparent that they are trying to hide the fact that it's in their dataset instead. They're still making use of the data though, just putting a system in place to try to hide their infringement.
Thanks for making this video. I checked out an AI kpop track on Udio, and some of the voices sounded like members of Blackpink. Their voices are pretty distinct.
Artists wanting fair compensation are not being greedy. The same people hurling that accusation turn a blind eye to the greed and hypocrisy of these AI companies who want to violate IP rights of others while holding onto their IP.
Found the one sane commenter
You can't sell your cake and still eat it. If I buy an artists CD and let a friend or AI listen to it then I have committed no crime. You don't still own what you sold. I can't sell you a car and then dictate where you can and can not drive, because I no longer own that car.
You are being greedy. Artists learn from art.
@@abram730 If you're so against the idea of Intellectual Property then why aren't you just as opposed to AI companies claiming their software as their IP?
If wanting to have the same rights over our creations as the AI companies do over theirs is greedy, then so be it. Very telling that you're hurling the greed accusations at a generally poor group of people and not the corporations being funded by millions of dollars and still growing.
@@muzakbling8233 Suno is charging $0.02 a song for full rights to the artist.
I'd point to those leading the lawsuit as those responsible for the artists poverty, as well as the gang of haters who said artists and content creator is not a job.
On Spotify creator only gets 6%.
They did a deal with the publishers and require you to have a publisher for example.
I'm saying that IP has gotten quite dystopian, and gave examples of real world objects being treated the same.
I hear Adobe now claims all works made with their software.
An artist can't be an artist without using the works of other to learn from.
Great work Jesse! just experimented with Udio, trying to get it to spit out a song in an Elvis style by saying “asking for a 1950s style song about blue silk shoes” and it totally spat something out with a different melody but ripping off the Beatles song lyrics to “Get Back”. The lyrics from Udio were: “Step Back, Step back, Step Back to where you once belonged”. I hope they get sued to oblivion!!
As bad as this sounds, one thing AI or these sites will not be able to do is innovate. It will always be based on what humans create. For example; when the first distorted electric guitar was recorded it sounded terrible to them, since all the recordings up until then were clean. Till one day someone said that it sounded good, even though there was no “data” to base their decision on. If there is no data to support it then these sites won’t be able to recreate it. So therefore it takes a human to innovate or come up with new ideas.
Great work Jesse! Thank you!!!
How long have the Top record producers been using this sort of Ai technology?
There is literally zero question they were trained on copyrighted tracks. The evidence is pretty damning here.
Don’t speak for us… we don’t all know and love. Record companies been screwing us and artist for decades… the chickens are finally coming home to roost
@@UNBIASEDMEDIA you’ve had a deal with a major label that went bad?
Surely the law will have something to say about this!
I wonder whether it's possible that both Udio and Suno will have a similar affect on the music industry as services like Napster and Limewire - they absolutely changed music, but ultimately artists adapted and found a way around the issues the services presented.
Fantastic video. Another great one, Jesse.
This "Human Creativity" part is the most annoying thing I've ever read... Thanks a lot for the video. It took me time to understand how AI draws... Now I would need to understand how it composes music :) especially without disclosing sources :)
A few issues:
With this one, a bit harder to prove. With the ai songs rains of castamere and ayy, you can tell they put the lyrics themselves. With rains, guy messed up verse and ayy had an intro that suno ai can't do, but you (a human) can manually do it yourself.
Also, ai pronounces the words depending if you put caps or not.
Also, suno from june and afterwards can use audio upload so it makes it harder to prove with a lot of new tracks especially when you have white boxes with red outlines blocking important things in suno to see how it was done.
A few more things need to be done to show it's using copywritten.
The first one was found by accident and doesn't contain any word from the producer tag.
@teko363 have you used it?
All the first samples are from version 3.0. Suno is currently on 3.5.
Also, not showing the whole lyrics on that first song.
I found the first song. Dude only shared that one song and that one post on reddit and that's it.
No way this is a fluke. The timbre of the sounds is too similar to be a case of someone caps locking the words
this is hero stuff, Jesse!
It's more like having the instructions for creating a copy. You don't have a "sample" of the sound anywhere, but the AI has learned the steps to take to create a close facsimile.
Down to the waveform
@@davidcuny7002 it’s not the “steps to take” as much as the exact frequencies to replicate, as demos like this prove
@@bunnyyyyyy398 It's likely that the training data consisted of audio frames dividing spectral frequencies into Mel-sized bands rather using "exact" frequencies. By "steps to take", I mean the neural network can output frames containing similar spectral curves that approximate the training data. The point is that "samples" aren't stored in the network, but rather the ability to construct an output that approximates the training data. By way of analogy, it's akin to having the recipe for baking a cake rather than storing an exact copy of a cake.
You just described making a copy of an audio file...
@@bjornskivids Is there evidence to support this claim? Something can sound the same without being the same waveform. Even if the harmonic content is identical, if the phases of those harmonics differ, the waveform will look very different. The EnCodec audio compression codec Udio uses (at least with Bark) seems to have trained the discriminator to a resolution of a "64-bins mel-spectrogram", and I see nothing the paper where they attempt to maintain phase.
Um its happening cuz users are probably uploading known songs cuz now have that option for others to get the outputs. The thing that doesn't make sense js if suno got so much money why not do it right get licenses etc wk with the artists and producers and then disclose their training data. And now in this situation they are in where do the users stand. Cuz soon as memberships payments ny users hit music industry was like oh no saturation and money cut. It's not fair to everyone all around
This was discovered before Suno started letting users upload their own audio.
The question that needs to be asked is why is it only @SyncMyMusic finding these materials. He's more than likely intentionally trying to sabotage Suno and Udio. Out of the thousands of people that utilise these platform he's the only one coming out with these results and claiming to have "sources"
@@tinagomez4418 ok I’ll help you out. I’ve been talking about AI on my channel for a year. A subscriber reached out to me with the Udio findings. That video led another subscriber to reach out to me with the Suno findings. When you talk about specific issues, the internet notices and interacts with you.
@@SyncMyMusic Help yourself and save your time explaining. That narrative doesn't prove my suspicion wrong.
Bro ive been using suno and honesty you can add that tag as "lyrics" then go back in and erase the lyrics so it doesn't appear stop being so gullible
Who cares. Let’s go Suno!
Regarding your "sampling on a spectrum" and "fully masked" - If it's anything like what we're seeing with the image generation, it's more likely that it's just filtered through someone lesser known, rather than being any sort of significant amalgamation/something unique. In the more niche visual subjects, it's incredibly obvious just how much the tech is just a melded or tuned collage because of how reliant it is on specific training data. Suno and Udio seem to be doing the Midjourney route, in that the "improved" models are just weighted even more heavily toward the source data (why the latter spits out near 1 to 1 movie stills now).
*Edit* Also, regarding the 50/50 split, it's worthwhile to remember that this technology isn't new. Diffusion has been around since 2014. The -only- change and what caused the recent explosion into use and public consciousness was the application of the new indiscriminate datasets.
This is interesting. Do you know if the visual artists are fighting back at all?
@@jmi_music Trying to! In the US there's a class action lawsuit that's going forward with multiple infringement claims (there's a website for it at "imagegeneratorlitigation" or see "Anderson vs Stability").
Last year we also funded a full-time lobbyist through the Concept Art Association. Though efforts this year there are stalled due to fundraising issues - a lot of people have already lost work and can't contribute anymore.
There's also multiple European orgs lobbying, such as EGAIR. Seen Brazilian and other countries with movements a well.
There's also now a new platform that artist have moved to due to scrapping concerns called "Cara".
In the end, it's going to come down to regulation and lawsuit results. And the court results in visuals might set precedent for other industries as the proceedings started earlier.
@@jmi_music Sorry. RUclips is eating my detailed response, but the short version is yes. Cases (which might set precedent for music, etc, as well), lobbying (regulations and opt-in approach), various orgs (US, Europe, Brazile, etc).
Sadly, you just convinced me to stop using both Suno and udio. I had a hoot with it and created a few rather liked tunes with it, but because of this I am a little upset with both of them and myself for diving in blindly. I do hope that everyone used to train the A.I. so get something from it..... Heheh, just last week I created something using Suno about this, and in my creation, I repeatedly say I hope the artists used to train the A.I. get rewarded well.
My "human resource" is just as FREE to use for inspiration as any OTHER piece of music out there is. Musicians have been 'influencing' music for years. When was the last time you heard an artist say to another artist "Hey!!! You're music sounds like mine! Did you actually EMULATE me without my permission"!???
Can you imagine, Eddie Van Halen hearing some young guy playing guitar and saying "Hey, did he just use the same whammy bar technique that I used in that one song WITHOUT MY PERMISSION"!!!???
People aren’t computers. Human creativity doesn’t work the way machines learn to plagiarize. Copyright law doesn’t exist to protect machines.
@@bunnyyyyyy398 Not supporting plagiarizing. Anyone (or anything) that plagiarizes should have to pay the artist. Plagiarizing means COPYING verbatim, not 'drawing influence from'.
@@bunnyyyyyy398 "Copyright law doesn’t exist to protect machines."jj
You're right! They exist to stop people from COPYING, DISTRIBUTING, and BROADCASTING, so if AI has not 'copied' or 'distributed' or 'broadcast' any works without permission, then there have been no laws broken. Copyright law also do not exist to stop people (or computers) from LEARNING from copyrighted works.
These cases will go nowhere. You can expect the same results that we saw with AI art, and AI writing. You can't copyright a 'feel' or a 'genre'.
@@rockandroll798 copyright law exists to promote art and culture by assigning exclusive rights to authors. Whether those rights extend to protecting your works from being used by AI companies for their own profit has not been decided by the courts. Your confidence betrays an underlying naïveté about the entire situation
@@rockandroll798generative AI isn’t “drawing influence from” in the way the phrase means (human behavior). It’s processing and analyzing patterns within data to generate responses based on learned algorithms.
In the context of music, AI can inadvertently create works that closely mimic or replicate portions of existing songs. There is no original inspiration or creative intent. Essentially, AI isn't creating something new from its own "influence" but rather reassembling elements of what it has been trained on, often without the nuanced originality that human artists bring to their work.
Unreal. Copyright exists for a reason 😅
The use of the word train is improper when describing AI music producing because that's not what it's doing, the operative word should be "recreate" because the new version doesn't have to train, and When you figure out what it is really doing then you will realize that you have no protection from it's capabilities!
50% Value from the Data-Input? I think way more. The artists shall get paid properly - also the unknown artists who offer diversity without having success on their own (might have some financial blessing out of a sudden). Remasticating 'main stream' artists might offer much less value for the AI-quality than other artists, that should also be considered. But destroying Udio would be a gigantic shame for mankind.
Suno has an upload function so users can upload their own audio samples and put them into their tracks. Suno will ask the user to confirm the audio is their own and not copyright material. So it's the user who is infringing the copyright by uploading, not Suno. Suno does not need to sample any copyrighted music to produce the music, that's not how it works.
It’s effectively spectral sampling. If you know how that works. The tags are being regenerated. Suno can’t hide behind “oh some we let some rando upload DJ Mustard track so we can do whatever we want with it “
@@bunnyyyyyy398 Producing a spectral sample is entirely different from creating an actual sample. Suno does not simply sample music and insert those samples into tracks. My point is, Suno would not replicate a producer's tag so precisely and then randomly place it into a song. In my opinion, these cases involve users taking samples from songs and incorporating them into their tracks using the upload audio function. It's peculiar that this is occurring now, just after the upload function has been released. We know a lot of people hate Suno and Udio, and we have to consider if they are doing it to discredit them.
@@newsjamgroove0 Nope, we didn't upload any audio to create these outputs (only prompted them). And this was discovered before Suno released their audio upload function.
@@SyncMyMusic
Yeah we'll just take your word for it...
😂😂😂
Is there any chance that Suno will get banned and our songs wil be trashed and forever lost? 😢
Yes unless you download them
can a technique of training be copyrighted? Can a technique of sampling be copyrighted? If not, then does someone using that technique have rights to anything of value other than the implementation and output of that technique? Is there any precedent for this type of intellectual property right? For example, PCM digital recording is not owned by anybody, but the implementation of that technique determines the copyright of companies that utilize it. Those companies do not have to pay the entity who came up with the technique of PCM recording, as far as I'm aware, unless some aspect of the rights owners original implementation legally requires their licensing of those methods. seems to me that the argument is being made that someone learning how to make a particular sound requires that person to pay the entity who owns the intellectual property right to that sound, regardless of whether or not the entity that is learning to make that sound is actually reproducing that sound . If the lawsuits were about the output it would be otherwise, but if it's about the training, is this not the argument that's being made?
@@aaronseckmanmusicif you swap PCM with Dolby Atmos in your scenario, then yes there is.
Also, that's not copyright, but a patent. Copyright is not the appropriate IP for that scenario.
You dude are behind! You prompt something, it will come thru!
How do we know you guys uploaded the audio of those tags and made Ai learn 'em?
Ai learns even as you prompt or load guide sounds...
So the companies won't win. Place a bet...
Ai gives you what you put into it, it will not give you a producers tag by a quote on quote mistake, this is bogus. Everyone on this earth learns from each other, stop being scared and teach Ai to be good, or it will turn out just as how we humans are.
You can see the prompt for that first track with the "CashMoneyAP" tag included in it. There's no prompt for those words anywhere.
@@SyncMyMusic
Another lie.
It's the title of the track.
😂
Suno reads that too.
Dude you're so dishonest.
The music industry simply like all other entertainment business should not be allowed to have copy rights on language and music which is a common universal human right like free speech. Low life producer and artists with no real talent owning sayings and with holding human expression. To hell with them, long live SUNO long live human rights, down with the music cartel!!!!
@@MRNoory-vt3nb this is easy to say when you haven’t made your living using the copyright system.
this one is ridiculously obvious lol thanks for bringing light to it
It’s not just the training, it’s the reproduction of the material that is violating copyright. The output is more like thousands of clips pasted together rather than the actually creation of new material.
what if it's physical modelling?
this is insane!
Keep up the good work
SO EASY TO COPY WHAT KIND OF MUSIC LIKE IF I GO ON YOUR THINGS.... AND USE THE WORDS YOU SAY... TO UPLOAD AND CREATE A USABLE SONG... ITS LIKE WHAT YOU DO TOO.... YOU COPY OTHER PEOPLES MUSIC LIKE YOUR SONGS ARE RIGHT FROM EARLY 90s songs.
Wow! Great work
It's really cool meta did get their licensing in order. Between both suno and Udio I have 100 unreleased tracks. I wonder how many users haven't released because of the copyright mess? 😮
my man is out here doing god's work. Thanks for fighting for all of us.
i think the music making community is jumping to conclusions. What if Suno and Udio are listening to all recorded work and recreating it using Physical Modelling? How can we say that this is illegal? We all have listened to music our entire lives and when we sit down to make it we are basically reproducing what we have heard through our only interpretation. are we committing copyright infringement when we do that?
No. Human brains are not generative AI models. They function entirely differently for different reasons with different outcomes.
@@bunnyyyyyy398 i didn't make the claim that the human brain and AI are the same. Don't forget we still don't actually know what these sites are doing, ripping the music and using stems or creating the sound itself with physical modelling synthesis. I'm asking if these websites are listening to all music, and then reproducing something like what they have heard through Physical modelling, is it actually copyright infringement? i don't think it is
@@BuddaB911 Why no AI fanboy gets that it's not the AI model on trial, but the CEOs of Suno and Udio? They illegally downloaded copyrighted music to train AI. Whatever training of AI has to create a local copy on their servers, so to go out from Spotify, RUclips etc, which is forbidden. The CEOs don't listen to the music, they aren't artists, yet they profit from artists' work. They just violate copyright in the very moment they download the music, which is supposed only to be listen to.
@@federicoaschieri My friend you are way out in space with this comment. how do you even know if any music was downloaded illegally? What if they bought the Vinyl/CD? what if they recorded it themselves? what if they have a system where the system listens to the source no matter where it's from directly? Do you know the CEO's have never made art or music? Don't we all who make art profit from engaging all our lives in other artist work for inspiration? Non of us have answers to most of these questions therefore you can't make the claim that anyone "Copyright" was violated until you know for sure if it's a copy or synthesis
@@BuddaB911 Transferring the tracks on the computer is always illegal, it doesn't matter if from Spotify, RUclips or a CD. Copyright, as you appear not to now, prevents unauthorized copies to be made. Whatever system they use must do that before training. Your law-fantasy arguments need no answers, because are illogical. As I said, *we* listen to music, we don't download it. Law says that the first is allowed, because it doesn't make perfect copies of the digital files, while the latter does. So those pirate CEOs violate copyright.
Well on a copyright search on most platforms this would be tagged as copyright and not allowed.
Great video Jesse. Thanks!
Thanks so much for all you doing to bring these details to light. A glorified sampler is my term for the tech being abused by these companies.
They didn't "leap frog" those models, they are using those models under the hood they just scraped better data
Historically, I've hated producer tags... It'd be ironic if they were, successfully, used to show the truth of what these 'AI' companies may or may not be doing...
Booooom! 💥
Good video :)
I ain't surprised that they trained off CR music, it is basically the hot potato with all AI models, they need a huge amount of data, and they can only get that from the internet.
I have no clue if the record labels can win this or if it will end up in some strange agreement between the two industries. But I do suspect that the AI companies will get out on top, simply because the potential of AI and especially AGI/ASI is worth more than the artistic industry put together. I don't think any country especially the US want to lose this race, it could be catastrophic.
I agree with you that these models are created based on CR material, but I don't see how you would distribute money to copyright holders? Because if this should have any real value they would have to keep paying them, which they obviously won't do. The moment the model and technology are "fully" created, the music or entertainment industry is in the same problems as now. So I do think it would be fair for them to do it, but whether it is a realistic solution I highly doubt.
This whole issue as I see it, is linked to the huge issue being talked about with AI in general which is UBI, and still experts are far from certain that will work and they currently have no solution to what effects AGI/ASI can have on the world and especially the economy and power, which this is basically what it is all about, call it music, images or information it essentially all comes down to those two keywords as I see it.
It's not illegal to use copyrighted music to 'train' AI. It's illegal to COPY or to BROADCAST music without permission. It's no different than a PERSON listening to a copyrighted work, and then making another work that sounds LIKE IT, but ISN'T it. Musicians have been 'learning' from other musicians and then using that knowledge to write other works for YEARS. It's called 'influence'. Musicians INFLUENCE other musicians. AI is simply being INFLUENCED by copyrighted works, it isn't 'distributing' them without permission. It isn't 'copying' them.
@@rockandroll798 I don't think this reflects the issue. As far as I know, a lot of the legal issues are about these AI companies training from copyrighted material without permission. If what you say is true, then Suno and Udio could just as well openly announce that they did this and there wouldn't really be any issue. I don't think either Suno or Udio have any interest or particular benefits from having their models being able to copy existing music. First of all, I don't think any users of these programs would find it very interesting to copy an already existing song, they want to create their own stuff, also Suno and Udio know that users can't use or release these songs if they are copies.
The problem as I see it, is whether or not training an AI can be compared to that of a human doing it. And I think there are arguments for and against.
For it not being an issue, is that these models still require human interaction and it somehow falls under fair use, since/if they don't directly store samples in the model, which I don't think they do if it works somewhat similarly to the image models. Jesse showed a lot of examples, but at least a good portion of them I think you would be able to make the argument that the user's intention was to get the AI to copy an artist, obviously it might happen by mistake, but at least from my experience, it is highly unlikely. Therefore I also think you could make the case that the users have some responsibility as well, just as you could copy a song on your piano, copy an image in Photoshop etc.
Against it being ok, is that clearly, the CR material is integrated into the models and Suno and Udio have created some programs capable of copying CR materials. And even if that is not their intention, then they are somewhat responsible for making the CR materials available to be "copied/generated" without the right to do so. As these AI technologies get better, you will probably be able to generate songs pretty much on the fly, which means that you won't even need something like Spotify etc. You just tell the AI to create music in the style of whatever artists you want. Obviously, the music or any artist industry is not interested in feeding these AI companies if their intention is to grab all the money.
So it depends on what is actually legal and what isn't, if it turns out that it is legal to train on CR materials, then it will most likely have a huge impact on the music/entertainment industry because people will copy things as they please or simply create their own stuff and companies will definitely start using this much more, they can save so much money and probably more importantly time. And if it turns out that it is not legal, I think it will kill most of these (US) AI companies and then people will turn to use free ones (which are completely without any limits) or maybe some foreign ones who don't care to much about CR.
AI is all about repackaging and representing OPC. (Other People's Content)
There's no doubt whatsoever they trained on copyrighted material. The only question is if the US legal system will hold up or not.
A lot of this comment section is high on copium. Claiming users are forcing the AI to commit copyright infringement. Day trip is not a copyrighted word. How exactly did the AI decide to put a stutter on it?? Of the millions of ways the word daytrip can be modified, it chose to use copyrighted material. The people who defend Suno or Udio are not artists, they've never worked hard to produce anything and find it easy to justify theft.