Is Udio Reproducing Copyrighted Songs? (Audio Examples)
Вставка
- Опубліковано 16 чер 2024
- #udio #udiomusic #aimusic #mariahcarey #whamlastchristmas #feliznavidad #copyright #thebeatles #paulmcartney #johnlennon
0:00 Intro
2:52 Eminem’s Lyrics?
5:05 Mariah Carey?
8:00 Wham! Song?
9:32 Feliz Navidad?
10:33 The Beatles
11:14 John Lennon’s Voice?
11:57 Paul McCartney’s Voice?
13:06 Beatles Recording?
14:33 3 Big Questions
16:25 “Fair Use” Gonna Fly?
19:35 Red Lobster’s AI Music
20:32 Our Future As Human Creators
24:45 Reframe To Reduce Anxiety
Referenced Weaver Beats Video: • AI Music Has a Dangero...
Wanna learn more about TV/Film Sync Licensing?
www.syncmymusic.com/
Thanks for watching all! Before you ask:"Can you share the prompting method please?!" just know that we won't be publicly sharing the method that allowed Udio to generate these tracks as we don't want to help Udio clean up their mess. That's their job and responsibility. Feel free to play around with their model for yourself and see what you can come up with!
Then Udio and those who own the rights to "Last Christmas" and "All I Want for Christmas" should sue the prompt writers for plagiarism. Udio's just the tool they used and isn't responsible, just as GM isn't held responsible for bank robbers using GM cars as getaway vehicles. I hope someone lets Udio, Warner-Chappell, Sony, and Andy Stone know.
I don't think there is a special prompt. I think you may have been misled. Did that user share with you that he added the lyrics to the custom section? By adding the lyrics, all you have to do is add a simple prompt like Christmas or male voice or female voice or yuletide, etc etc and you will get everything that you just clicked on in the premiere. In the custom section, it override's the systems effort to create lyrics and uses those. Now how this user could use this to his advantage is create full songs with these and find a distributor who allows for Cover Songs and distribute the songs. It's a very good chance that between Thanksgiving and Christmas some, if not all, of those songs will get some playtime. I believe this was/is the main "business model" for Udio. This gives fans of X the opportunity to create new songs with their favorite lyric.
Again, nothing illegal.
Maybe slightly immoral but the power is still in the hands of THE FAN and THE USER and not the big labels.
And based on the Terms of Service, there is nothing these big labels can do about it. (And this is the real story)
Great video btw
and for clarity, to do what they done
1. Find your favorite song's lyrics with a simple google search
2. copy the lyrics
3. go to Udio, click on create and under the "custom" section paste the lyrics, hit create
in 40 seconds you will get a 30 second AI cover of your favorite song
to add the rest,
follow 1 and 2 for the missing sections,
come back to udio, select the "view song" and you will do "inpainting to add the missing section at the end. Continue until the song is complete.
@@VelvetHammerBooking Nope doesn't work. Moderation says try again with different lyrics.
@@JamesJones-zt2yxsure, if they paid for the training data
I love the optimism of the thought "I wonder *if* they trained on copyrighted music, and *if* they licensed it"
Of course they did train on copyrighted material, of course they didn't license it.
And that's why I wanted your comments! 😎
@@SyncMyMusic I understand ;) I'm ambivalent on your last point. On the one hand I agree that things like Udio can be a motivator. I personally find what it does offensively bad, and it makes me want to make music just to show that it can't do what humans do. I wouldn't go as far as to say it's here _for_ us though.
@@seban678 What if it were to use tools like this to A/B compare what we create vs. what AI can create? I've tried this and found that it's helped me to expand my own possibilities of my own writing.
@@SyncMyMusic Yeah, possibly, I hadn't considered that. It still seems like the whole endeavor is involving a lot of time, energy and data-mining for a minimal potential gain, and an astronomical potential loss. Thank you for your perspective though!
They have no trillion dollars to license every song they used to train their model...
Excellent work, Jesse. Thank you for the reframe especially.
Recently I heard someone say, "If these platforms don't let you use artists' names, it's not because they're good people, but it's because they want to prevent you from finding out what data they've trained their models with." 😉
That’s been my hunch as well.
Totally agree...what we need is a program that can scan the output of any Udio generated sound and connect the audio they are using to the original tracks...kind of like Shazam but more complicated....use computers to call out computers.
I think it's me.They could just not allow us to ask for artist's styles like suno is doing. But it can be someone else, but i"m going hard on spreading this simple implied fact.
@@SyncMyMusic The fair use argument is for using the copyrighted material itself, for instance reviewing a movie or a song, or parodying a movie or a song ect... The training argument is completely different, there is no law that states that content that exists cannot be used as a basis for training. Never has anyone been sued for using copyrighted material to train themselves on.
If artists and musicians don't have to ask for consent or pay licensing fees to use copyrighted material for their training, then why do programmers have to do it to train the programs they create? Why are artists and musicians exempt but not programmers and software developers?
If that's true, udio's doing a lousy job of prevention. On my published songs, my requests for "in the style of* are still visible, as is the more general description udio used instead..
Awesome video, @SyncMyMusic
Prompt: "Create a song that explains how Udio is about to be sued by every musical copyright holder."
Udio: 🎶"Sue, sue, sudio!" 🎶
Under rated comment lol
😂😂😂
🤣
Absolutely brilliant
@@gameforyourhealth yep 😂
Thank you for this, shared to all my social medias ♥
Thanks Jesse!! Awesome Video 💪💪
GREAT vid!!! Fascinating and mind-blowing Beatles renders
You have asked some interesting questions!...good job.
This is bizarre. I just posted a link to an example of Beatles song generation, and the video is now modified and the Beatles likeness and Beatles lyrics has been removed, but the video URL is still the same. That is one of the strangest things I've ever seen. It was "Chase The Sun" with both music and lyrics almost identical to Follow The Sun, except now it's suddenly a different song.
🤯
Whhhhhhaaaaaattttt?!
Maybe they saw this video and adjusted their system. But proofs won't go away on the court. they're really evilish.
All of these UA-cam videos are very good evidence
URL is still the same? 😮 Wth
I remember when the big 4 went as far as arresting people for sharing crap quality mp3s online.
Hi Jamie just joining the channel. You're absolutely right about us humans continuing to create music.. I've been working on a solution for years and we're almost there
In udio you're not allowed to upload lyrics if you don't own the copyright, so that's what happens if you do it
Thanks bro. This is needed.
If they didn't just outright train it on copyright music... How much do you want to bet that they trained it on music from youtube? Probably picked up a lot of covers of these songs as well as original tracks that escaped the content blocking, even if they excluded the official ones.
That wasn't John, it was PAUL on both the first 2! Close though on the first one, more raspy Paul though (think Helter Skelter).
thank you so much for this
Nice quote at the end there, thanks.
The best way to use udio,is to type your own lyrics, and add an acapella tag so u get vocals that sound original then u can just make a beat around those vocals
This is a tool, these dorks aren’t using the tools properly…everytime they keep PURPOSELY trying to INFRINGE. They arent doing scientific diligence by creating Original songs and then testing against HOW ORIGINAL the songs are to see if outside of the original lyrics IF music is original.
Great video!!!
The last phrase was very positive. Congratulations for being able to see that bright side despite everthing's happening. I feel that... Maybe... People will claim for more and more organic and custom material. Considering that AI is a soul-less music provider. Is too soon to know what will happen... Meanwhile, I'm backing to my roots, playing my old friend, the acoustic guitar. Which I always used as a therapy. (Btw, that's a great perspective to see the process of making music, as a therapy). I wish you all success. Greetings from Brazil.
A very good and thought-provoking video. These are very interesting times for all in the music industry, and my personal thoughts are that artists who produce very 'same-sounding' work, will feel most under threat. So those who push the envelope and continue to evolve their sound, will stay ahead of the 'learning'. It will push us to take more risks as you say, and for what can be a very safe and static landscape, I think that can only be a good thing, right? Thanks!
I prompted 90’s, New York, boom bap and it gave me vocals that had the exact tone, cadence, flow and inflections of Fat Joe, another sounded like Ghost-face (but a bit more like Action Bronson), and one that was for sure Raekwon.
I don’t know how it could be that similar without the models training in their catalogs.
An absolute fantastic research job here! This is a super meaningful analysis and needs to be shared widely
Don't you just love AI generated comments about AI?
Great video Jesse. Theres no denying now where they get their training from. upt o courts now I guess
Yeah like a kid trying to play these songs on piano. There are even licensed sheet music products that contain somev if these songs.
And frankly when you turn the radio on, every other song is now a 20-30 year old melody.
Copying music performance is not considered theft in any context. Commercial gain is, if it is not licensed. (Reminds me of that clip where a F1 driver is punished for singing a song he got stuck in his head. In no context is this punishable BUT he did it on live broadcast and the people responsible for the F1 events didn't want to risk getting sued so they took measures to stop him from doing that. Also live on tv)
As a studio engineer, composer and studio owner with over 40 years of experience, I have done thorough research with Suno and Udio. I loaded a studio mix into Udio, experimented with it, then returned to the mixing desk computer to compare the files. Although the composition was reasonable, the quality of the data was poor. I noticed a significant lack of high frequencies and dynamic range in the Udio-arranged piece. It serves more as a tool for composers and hobbyists to make music. However, the files are not suitable for production and must be re-recorded.
I just tried creating a song by copying existing lyrics from a semi-popular song (57 million youtube views) into the custom lyrics field. It's definitely copyrighted, published in 2015 and is published by the productionmusic label ExtremeMusic. They are well established and successful. Since productiomusic would be under threat too I doubt they would let AI train on their music.
I entered just about 20% of the overall lyrics. When the songs were about to be generated I got a "moderation error" saying "Your credits for this generation have been refunded. Please try modifying your prompt or lyrics and try again." My prompt was just "neo-classical epic emotional" So that can't be the issue.
It's insane but this suggests to me that they have stored those lyrics in some database and stopped the creation of the song as the algorhythm recognized them. But why would they store those lyrics? Just because they care deeply about copyright and don't want anyone using lyrics from an existing song? Did they build a dedicated system that randomly stored a ton of lyrics from existing songs, even less "famous" ones? Or... did they store the lyrics in a system that is used by AI to generate songs and they just try to prevent anyone from finding out with this trick?
Haha even Shazam detected that as the originals!
Crazy!
@@BenCaesar You know what's funny when I switch my fan off it doesn't recognize it. So with background noise it picks it up, but with a clear recording it thinks it's different enough.
@@luc.schurgersno way that’s even crazier !!!😂
Oh that's crazy which one?
Was Shazam "trained" on the copyrighted music?
Love the reframe 💜
Nice outro conclusion!!!👍🎶
Okay, the Christmas songs and such were using actually lyrics "cleverly" inputed as prompts, so the results (surprise) sound kind of like, or very much like published songs with the same lyrics. Of course, this can happen. The thing is, it is against Udio's policy and copyright law to use those lyrics. That is on the person who used them as a prompt, and you even acknowledge in this video that Udio will remove such songs. You and that other person moved them to another site to "protect" them so you could use them to take shots at Udio. Now, being a Beatles fan, I recognize that there is a soundalike quality to those (important distinction, here) Beatle-esque sounding tunes. The examples used seem to be original lyrics, unlike the Mariah Carey example and such. So, while the song may "sound" like a Beatles song, it is not one, and the singer is not the actual Beatles voice, trough it sounds similar. It just seems that you (or the unnamed associate) are being purposefully misleading, here. Again, Udio does not allow copyright lyrics, unless it's the human creator using them, and sound-a-like vocals and musical style parodies or homages that aren't actually taking the musci from a copyrighted song are not violating copyright. Were those Christmas tunes randomly created by the A.I., or "coaxed" by your pal using existing copyrighted lyrics? Where is your own transparency, here? Why not explain this process you and your pal used to expose the supposed infraction?
Personally I would like to believe AI music will allow some talented people to break into a music career where normally they wouldn’t have had a chance. While this is also opening the floodgates to a bunch of noise, there are gonna be those songs that just rise up.
AI is a tool. I personally don't care if they learn from all music a normal person could listen and train from. If AI needs to pay to learn so does every person who played stairway to heaven in the music store or while learning.
I hope that AI will destroy most of bland and sameish "hits" that are on the mainstream radio stations and are boosted by algorithms. However, I feel like when it become even remotely dangerous for the labels, it will get sued into oblivion.
@@artephank thats not going to happen ,is really naive to think that music generators will b3 destroyed because they are not simply music generators they are an all pourpose AI models that have the property to generate music or text or images as long as you give them the data , to think that could happen is akin to thinking that just because you destroyed one CD that your player plays you have actually destroyed the player , very soon entire AI models will fit in a cell phone memory capable of generating music with out the need of internet cloud connection you could sue one company and shut it down hi about the millions of people that will have their own private AI generators ? That will be operating in peer to peer networks or even the dark web exchanging music files by the millions ? Eventually it will just become the new normal ,trust me when i say in the context of AI music production is not longer very relevant when now people expect a full blown production with music and video and soon even virtual full inmersive reality with music just as one part of it ,this AI will produce
If you have to use AI you were never talented
@@artephank all AI will ever do is create bland and same ish music how can you people be this pea brained
Great video, congrats.
Hey Jesse, I've been a rebel since the beginning of my intro to sync licensing. I rarely used reference tracks and wrote what I liked to write. so, after almost 10 years struggling with this a lot of my music is getting the attention of a few libraries that didn't want anything a few years ago. I said from the very start the old school music will be back because it is good and not like the junk that is out there today. AI can kiss my behind since you have to pick your battles and fight them as they come. My strategy is not changing. That's my 2 cents, take it for what its worth.
Dropped a bomb here, man - great work 👏👏 Music lawyers are gonna be busy
No doubt copyright problem are coming soon but! I think IA open for us the opporunity to personalize our music, new sounds, real emotions and being original. This bring us to a new way as a composer without having to using IA.
Great insight ' thanks
Your video is *amazing*! Especially with the advice in the end! ;-) That's how I see it too - this is happening only to *awake* the Music Industry and the real Creators of Music, to get themselves in hands and start *creating* again, as it was before the internet!
SOLID GOLD!! somebody's got some 'splainin' to do
also the 80s was a decade of wonderful creativity and inovation.
Really well done, Jesse. I've been experimenting with both Udio and Suno and admittedly blown away by the results. That said, I really wonder how they "gather" the material to train their respective models. Chat GPT, Dall-E etc. are all facing multiple lawsuits for "scraping" copyrighted material from multiple sources.
As a musician, I wonder if these AI generated songs will make it into streaming platforms, commercial radio, Billboard chart etc. and if so, are they flagged or watermarked as being AI generated? (suspect most people won't care - especially if it's a good song). ...Interesting times!
You can bet spotify etc are already being flooded with AI generated songs
Certainly interesting times! And yes once AI music is indistinguishable from human-made music, all bets are off in terms of how things can be properly labeled online. I certainly want a clear labeling system so we know that's made by humans, but I'm not aware of the clear answer to that yet. Thanks for the comment!
@@SyncMyMusic Thanks, Jesse. After watching your great piece, I watched a video about the legality, copyright and ownership/ethics of AI-generated music by "Top Music Attorney" (Suspect you're aware) and she brings up some very interesting points about the ownership of generated tracks. Curious how PROs like BMI, ASCAP, SOCAN etc. will deal with the inevitable onslaught of AI music. A bit of a nightmare but people will inevitably try to beat the system.
@@west4coast77 yep I had her on my channel recently!
A voice cannot be copyrighted. According to the legal decision in Midler v. Ford Motor Co.: “A voice is as distinctive and personal as a face. The human voice is one of the most palpable ways identity is manifested.” This ruling did not impact general copyright, but its subsequent legal interpretations means that while a recording of a voice may be copyrighted, a voice itself may not.
As for the lyrics. Udio did not just create those lyrics. Whoever prompted the song just included them in the prompt.
He's right. They cheated the system. used incorrect spellings when he prompted manually. done it myself many times. Udio can be cheated but suno it is harder to do.
I mean... two people can have the same voice. idk how that would work, legally speaking, if you could actually copyright.
@@DjGolucky Oh wow, if that is true then lawyers can't use that, Udio does block full verse then. If Udio shows it blocked lyrics, names, etc.. and they also paid to "listen" then they have solid ground. It all depends on if they paid to listen, and honestly it would be pretty hard to prove the did not.
@@TheFeedRocket udio will block copyrighted songs. But all you have to do is be creative. So instead of spelling "its a beautiful day", spell it "iz aye butifull daye". Get it!
"voice may be copyrighted, a voice itself may not." - but the thing is they used copyrighted recording to recreate this voice. Is like cutting painting into million pieces, copy them on xerox and then connect them again into the somewhat similar painting. This is just derivative work, in addition to the original. Neural networks do pretty much same thing.
What do you say to a person that says they don't have a problem with this any more than they do with you covering a favorite song? The violation is in the USE of the generative production, not in the production itself. So generating it, shouldn't be illegal. But if I then used it in my film or commercial, just like if I used your cover of copyright material in my film or commercial, THEN i would be in violation and subject to litigation. I'm just asking.
Correct, me singing a Mariah Carey song at my private family Christmas gathering (without paying her) might technically be copyright infringement, but it's not enough to warrant any lawsuits. It's a good question you bring up and likely a pro-profit scenario would have to be in play for a lawsuit to succeed against a company like Udio. I'm not a lawyer, of course!
(imo) If I am a large online business and I host copyrighted music on my website, that is a copyright violation. In order to prevent illegal hosting of existing IP, music generation models must be incapable of generating any material that would infringe, or they must be locally hosted. But the distribution of a locally hosted model capable of infringing on existing IP is still legally murky. Also, detection of which generated material infringes and which generated material does not is difficult to automate.
Personally, I think that for every IP holder that matters (large artists and labels) deals will be cut with AI companies and everyone else will get fucked (me included).
If you want to generate it you can feel free to play it yourself. AI is taking and using a copywriter production. You’re not doing anything, just the AI stealing.
Pandora's box has been opened, you're fired. If Udio or Suno are denied this, some other company will take care of it. This is not just about music, the world is going to change. Adapt or die
Okay, hi! I'm Lynn Cole. New subscriber.
I'm an experimental artist. Suno got me hooked on making music, so I bought a keyboard and a daw.
Now I'm an experimental musician and producer too. I have two kinds of music I make these days, 1. Sweary country music about drugs and toxic love affairs. 2. Trying and failing to be Frank Zappa. If you saw my sound cloud, you would understand. My lyric videos are okay, but SC's really the place to track my journey as a songwriter.
I have some thoughts here.
First and foremost, you have the right approach and attitude. You do. That's because you're an artist, and you get it.
Next, Red Lobster used Suno, during a period of time where the model was REALLY fried, and you can hear it in the tracks. If you pull it up with a spectral, you'll see that the mid pass is all messed up. That's suno v3p1. They're on either p2 or p3 right now. V4 is coming, or should be. If it were me, and I was suno, it would be coming a month ago.
That mixtape was awful though. It's more than cringe worthy, and everyone who's looked at it and said, "wow, I hate it." You're right. It's a zero effort campaign, and it shows. You noticed it, their customers noticed it, everybody noticed it. Does that mean an end to sync fees? Dude, I dunno, how scared are you that the people who buy your work have no taste?
I feel like, if you're ever worried about fully ai generated music. Just go to suno or udio and tell it to generate an entire song with no human intervention, editing, or songwriting, and see what comes out. I got news for you. If you're impressed by what comes out lyrically, you have absolutely no business calling yourself a songwriter. None. Just go home, take your neon light cliches, and call it a day. You did good, you'll be replaced.
If you're not, congratulations! You're probably at least adequate, and don't suck at your job! Achievement reached.
As someone who tells herself she's making serious art, I want the legal consensus in the US to be reached on this. Right now, I'm only famous in China, and that sphere of markets. I would like to be taken more seriously by the music industry, and the only way to achieve that is by legal certainty. Regardless of the outcome in the courts, I support it, and look forward to what happens. No matter what, it's going to be good for everybody.
Also, please post hashes for those tracks so I can take a look at them.
Would love to share this with Udio support.
Yeah, I discovered fairly early on that melodies and chord structures of at least Christmas music seemed to hold pretty strong coherence. I generated an orchestral piece of Hark the Herald Angels Sing. I didn't think to check copyrighted music though. I was just surprised to hear the A.I. just straight up generate the actual melody and chords. I figured Udio knew it was was public domain so they allowed it to happen.
Just my own speculation, but I'm wondering if they were working on a way to do generative cover songs with a branch of their model that retained melody, lyrics and chord structure and they were experimenting with Christmas music, so they filtered the music they were scraping from which had key words like "Christmas" or "Festive" through this 'cover music' variation of their A. I. model.
But IDK how it all works. That's total speculation.
great video, entertaining and also this video can finish in court as evidence
are you sure you did not write the lyrics in when you created the loop
That is probably exactly how it was done; you can see "Amazing Grace" which is the #3 staff pick on their own website. Interesting that the video creator chose to omit this info
@@MaxLohMusic seems like false information to sabatage, from an idiot.maybe jealousy rival company
@@mattpearson1914 I wouldn't go that far; it's an omitted detail rather than a lie, and he's a musician, not a rival company
Thanks for sharing ! I make music now for myself and not for sync anymore. I'm depressed to see this things. Udio en the others are gonna make a lot of money.
Is that Udio interface? It looks a little different. But thanks for the video, I like it.
I did the same thing. Asked for a crooner sound like "Frank Sanatra" it gave me that message saying it can't reproduce Frank Sanatras voice. It said it replaced Sanatra. But it did end up sounding exactly like Frank S.
Does typing sanatra three times summon sanatra or something and why is it a bad thing to do so.
Btw, Sinatra*
As a Reaper user, I recently installed a Python script called 'Composer Assistant' on my Reaper DAW. This script harnesses AI neural-based technology to generate various versions of the original melody in MIDI format. Interestingly, a similar process seems to be unfolding here, albeit with audio. Given that every song can be deconstructed into its individual instrument parts, it becomes quite straightforward to craft alternative versions. Simply reassemble these parts, and presto! However, it's crucial to recognise that this practice is illegal. Despite this, the company behind such tools may attempt to evade legal repercussions, perhaps due to the somewhat lawless nature of the current technological landscape.
" lawless nature of the current " - there is just not enough money in that at the moment to bother. Look what happened with sampling. Until '90 nobody bother to sue. It seemed "legal" to use samples of other artists in songs. Now producers publishing samples without clearing are sued into oblivion. There are tools that can spot even less than second of recording used. I bet similar thing will happen hear. I think we should just use it for fun till it lasts:)
It's a whole new ball game when entire songs can be manipulated. It wouldn't be surprising if legal action follows, leading to new laws for protection. The pace of technological advancement often prompts legal adjustments. It'll be interesting to see how this unfolds in the near future
Dude the All I Want for Christmas is crazy :D But I thought the person who generates the track prompts the lyrics?
The more videos I see like this the more excited I am about the type of music that I care about having a renaissance. Unless we’re willing to allow culture to die, creators will keep creating and making and combining new sounds. Thanks Jesse for leading the charge.
I'll be interested to hear what the libraries think and how they plan to deal with AI. With Google and BMI getting cozy this will get interesting, to say the least. Great video by the way ya had me laughing big time.
Thanks! Yes I plan to talk to all the Library CEOs at 'Sync Up' this year about their thoughts on all this.
WTF Udio? I hope they get sued accordingly. Thanks for the amazing content, Jesse.
@@AILEYMUSIC You shouldnt be able to do this though you spanner. Plus you can copyright lyrics and Udio is clearly using them.
@@AILEYMUSIC Point being... you shouldnt have the option to be able to clone songs this easily.
@@AILEYMUSIC I don't think it would be possible to prompt the AI to generate the exact same melody harmony and style of a specific sound, like here we heard the identical songs like Last Christmas and some others. No doubt this AI is trained on the copyrighted songs illegally or because of the lack of law on AI generated arts at the moment.
Anyone can record a cover version of practically any song they want as long as they to pay the proper royalties. All Udio needs to do here is track songs that sound like existing songs and pay the required royalties. Or, am I missing something?
Actually, if they’re using the artist’s voices, that could be extremely problematic. So, I may be answering my own question. 😂
Honestly, it's kinda cool that your friend was able to do this.
I've tried to do this on udio. It's difficult to impossible to figure out. Also, I'm really curious how your friend got around the moderation system to make this happen.
You're making it sound like the standard use case, when what you're clearly talking about is a hack. Please share the details of this exploit.
Thanks.
put "style beatles" it generates, some similar ones appeared for me too, and paul's voice is clear haha tries a few times it should appear for you too
There will be a renaissance of original music that is authentic. For those of us that are truly musically inclined, there will be a future. Focus on finding you own original sound.
Genius video! Thank you Jesse. Dedicated to all thinking that AI works like human intelligence. I am happy to see confirmed what I have guessed in a video a few days ago: AI music is basically “photoshop” on steroids. The way the model works is stealing chunks of real music, changing lyrics, modifying a note here and there, and merging the chunks together. I found several examples as well. And all that on top of the fact that the theft of training data is not fair use at all. Thanks for this very clear contribution to spread correct information. We need that because people are deceived by the AI propaganda. Let’s continue to inform people!
That doesn’t confirm that. If anything, human musicians can reproduce songs they know when asked (though not in the same voice obviously). AIs memorize some stuff and can be prompted to reproduce, but it has to be done in a way where the aim is to make it reproduce. And it will not be a 100% reproduction, because AIs are not some amazing compression algorithms, but yes, they do memorize some things to some extent and can reproduce them when forced or randomly, though in that case with a pretty low chance. Same with image generation AIs. Anyhow, who says that it works like human intelligence?
And i’ve said it before, it’s not in the interest of any country to stop this, because not every country will stop it and whoever does will put themselves at a disadvantage and a serious one at that. (But we have seen a lot of things being done by countries against their own interests and security, so who knows, maybe…).
As for musicians making new stuff now all of a sudden, that’s a nice notion, but it’s not like new, completely novel music styles popped up all the time during history. They did, but they died, because noone cared about them. It always took a genius to come up with something new that got people excited. But geniuses don’t grow on trees, it will be very rare that someone makes it like that, those that get hit by AI will get hit by AI, the overwhelming majority of them are just creative people, but not geniuses.
And I’m not against you, I am one of you, I also stressed out over this, because music is my life, I played and made music all my life and when udio came out I just stopped, like what’s the point anymore? But then I created a concept album using udio just for me, haven’t shared it with anyone, but that’s what I’m listening to now most of the time. And why? Because I wrote the lyrics for it and I made it in 30 second parts and chose what I liked, so the end products are exactly how I like music. I had more than a hundred generations for a 4 minutes song and threw most of those out. And this gave me a different perspective. Yes, as a musician this physically hurts. But as someone who likes to listen to music it’s a game changer, personal music for everyone. Exactly how people like it. Exactly. Not in the way of yeah it’s a cool song you made, I like some parts of it, all in all I like to listen to it, but it’s a 100% your taste and only 60% mine. People won’t have to rely on someone else’s taste in music anymore, they can make any kind of music exactly as they want to. And though at first I didn’t think of it this way, now I see quite a bit of repetitiveness in udio, it has a style even if it can create in many genres. Just like you can tell most of the time that an image was made by midjourney, now I can kinda tell if a music is udio and so now I believe good music will stay being good music. And though right now, at this moment my favorite album is the one I made with udio for myself, i still like to listen to other music and I still look forward to new releases by my favorite bands and I still love to play Chopin. But I don’t make my living from music and if I did I’d also feel threatened, but I’m not as worried for music as art anymore as I used to be.
@@eriong.7446 Lol, if humans make covers they have to pay. When it is to your advantage you claim AI is like humans, but when there is to pay, then it is a computer without credit card 😁 And your words unfortunately are useless, because even if the model were more creative, the theft of training data, as Jesse well explained, is not fair use, so the model is illegal.
yep it need different stuff to mix together to get illusion of creativity. it why generative AI can not train of other AI. voice cloners likely can train off other clones because it just copying.
@@federicoaschieri 1, i have a hard time making myself reply to a comment that starts with “lol”.
2, to my advantage? What is my advantage?
3, have you even read what I wrote? It doesn’t seem so. Where have i claimed it’s like a human? I specifically asked “who says it’s like human intelligence?” I only said that what you claim as proof that it’s unlike human intelligence is actually more of a proof of it being like human intelligence than of the opposite. It has nothing to do with royalties or credit cards or whatever. I don’t even know why you brought that up as if royalties would stop people’s ability to create covers. We are talking about ability not getting paid for it or whatever.
4 you can repeat your theft claim, but I argued why I(!) am not sure it will stand in court, even if morally correct and you addressed none of that just repeated the original claim.
@@spark300c yeah, like chatgpt has all those pesky letters it can mix together. Have you guys tried it? It works with melodic lines in harmonies. It repeats sections with different orchestrations, or the same melodies with different harmonies or adds a different instrument playing counterpoint to the original in repeat and stuff like this. Sure it needs to get the sounds somewhere and it copies it, but at that point where you can say this 0.00000000000000000001ms is from that song mixed with those 4000 other songs and the next 0.000000000000000001 ms is from those 5000 songs then what’s the point? Saying “it changes the lyrics” is like saying “it only changes the molecules to turn that water into wine, no biggie”. You can’t just change the lyrics of someone’s singing, you need to regenerate the whole thing… how do you “just change the lyrics”?
🤔 💭 uncanny. Udio got a problem. But when it comes to laws and changing laws big business can buy off court officials to get laws passed in their favor
Is there a way to prompt the chord structure. Each of these always generate two songs, but none of my udio or Suno sound alike.. what can I do to control the chorus progression?
'Why is this happening for me' noted and locked in.. Thanks.
That's actually the most part of this video IMO.
'Training' seems suspiciously close to 'stealing'.
Do you consider it stealing when you listen to a song and try to learn in on the piano? Or even play it with your hobby band in some garage?
This is training.
Now turning what you learned into a commercial product is another story. There is a difference.
@@borstenpinselthis is really silly. Covering a track at home is not the same as just completely recreating it, lol.
@@BrofUJu yes it is. The difference is only in the access of the public and the commercial nature.
@@borstenpinsel okay? This is a commercial product. If you post a Beatles song onto UA-cam it's getting pinged for copyright. Should be the same here. If I play a Beatles song in my bedroom on my guitar no one is pinging it for copyright, lol
@@BrofUJu yes but the comment was about the training of the AI... but I see you understand my point. It is NOT stealing to _train_ on replicating songs.
Good for you! As someone who has been robbed of untold revenue, I am interested to see where this goes. As composers/songwriters, we should hold to account our elected officials in this matter.
Personally, I have been robbed of God Only Knows how much income BEFORE AI. Yet there's a RICH irony in how all media outlets punish, with copyright infringement, anyone who uses copyrighted material via public posts.
Great job! Please keep up the good work. To composers/songwriters, NEVER STOP CREATING. We will eventually win this war. AI will NEVER stop, but it will NEVER stop us if we keep doing what we do!
That "Last Christmas" song though naturally falls into those chords that bring meaning to the lyrics. I, myself, as a composer have found certain lyrics I create and write music for ends up sounding very similar to other songs on the same subject. I made a song in the early 2000's and was accused of copying a song that had just come out that I hadn't even heard yet. The title is even the same! (Ingrid-Tu M'as Promis vs. Fresh Nelly Tu m'as promis) I began writing that song long before their release date. I'm sure this is what's happening in part but yes, there is an influence present for sure!
I love your point about aiming to break boundaries with our music. Essentially, that's what the Beatles did in the sixties and if AI would have been around then, it would have had to catch up with the new trends. So indeed, AI is only generating the music of the past. Humans are the real creators.
I knew it, too good to be true!
It doesn't sound like it's "generating" music, but instead using modified audio bits from the original recording.
THAT’s one of the very important questions I have as well!
Nah, that's not exactly how it works although it can feel that way. If you've ever seen how image generators work it's a similar phenomenon from what I can see, where you'll see pieces of watermarks show up. They don't literally take pieces of the training material, they generate content from a random seed by calculating probabilities if what something should be based on tags and Metadata provided by the user and on the training materials
@@dinierto i'd still argue it's a form of sampling. Even if you recreate a sample (or "generate" as you call it) you'd still infringe on copyrights of the songwriter.
@@diniertoDisguised form of sampling.
I love the "bold, inventive, originale and more creative" approche as a strategy to compete with AI generative music models in the future. I can't wait to learn how to apply that approche to production music. Do you think that could work for music licensing?
It's a MUST for sync licensing.
well done !
There is a man under the table pulling levers and your being told you’re seeing the future.
What? you may dislike the moral aspect of this, but it is the future lol
@@Zer0Spinn uh ok I guess we’ll have to wait and see. If you don’t get the reference I’m making then you likely can’t see that you’re being fooled. Clearly can’t see I should say.
These songs (udio)It sounds exactly the same wit the originals songs
Reminds me of the early days of image generation. If you prompted them in just the right way, you could get them to sorta recreate certain very very famous images. But as the training data expanded and diversified, that problem went away.
I was hoping you did an example of The Beatles and thank you for doing so. The John Lennon likeness wasn't much like him but Paul's voice was. And the Drive My Car beeps - omg. Despite them being in a different key, you would have to analyse them in a frequency spectrum analyser to find enough difference to avoid copyright infringement. But gee, gimme a break there. Too much alike.
AI really brings out the fuddy-duddy in people.
Welp, here comes yet another music industry shake up. This should be interesting
We’re certainly not low on interesting times!
Of course the model contains copyrighted music.
Now, the Copyright infringement happens when you put the same lyrics and trick the prompt to get the artist anyway.
It's up to the user, like with every tool, you are responsible of how you use it.
Also, with the inpaint feature, you can torture a song until it makes the exact chord progression you want.
Very interesting video. Much enjoyed your opening to this debate. As a time served (40+ years) producer of music here are my thoughts -
1) Licensing will come when the matter is tested, either in court or as a result of a threat of court.
2) If Udio (or any of the others) have any brain cells they will either - already have the licensing sorted, or are anticipating the need for licensing.
3) Fair use doesn't apply, for the reasons you outlined.
4) Udio in particular IS capable of original thought. It CAN take a relatively complex prompt and produce things that have not been heard before (at least by me). The four most recent tracks on my channel illustrate this.
5) Musicians who currently make a living from selling their music as a commodity, for sync purposes or for corporate use, are probably doomed. If you are a creative original, and you don't rely on your music to pay the bills you'll probably be OK. Indeed, you might find your work even more sought after.
6) This one's just an opinion, but creative people must embrace the AI and use it in a way that helps them. If they don't they risk being left behind. I'm thinking of AI applications within the studio/DAW environment.
Great video, and very interesting times.
Shouldn’t you show the prompts? You can enter custom lyrics. It would be valuable to show your prompts for the Christmas music
And give Udio the fix for their bug? Nah, that's their job.
@@SyncMyMusicfair, but then we only get half the story. You state that it was able to recreate “lyric for lyric”, but it’s possible that the lyrics were inputted specifically by the user. Just saying for transparency.
Agreed more transparency please
That's right. With Udio, you can add custom lyrics. If the user typed in the lyrics to a copyrighted song or very similar ones, that's not Udio, that's the user. Actually, I would never do what that user did because it's a violation of Udio's terms of use. Also, if the copyright owners go after anyone, it would be him. The generator only did what he told it to do.
That doesn't explain the note for note melody.
I think Udio should have the right to let the AI listen to lots and any music and learn, just like humans do. I do not , if I compose a song, try to find , who I have been influenced by and then ask their permission to use a short sequence of chords or feel, or style and ask how much I should pay to use that idea or chord progression or for example ABBA -like mixing method or Hollywood Strings sound. We all learn everything we know from the world around us. There are laws that protect our copyrights and those laws define, how much similarity a panting, a book or or a score can have without it infringing that copyrighted or creative work. God luck, Udio! You help us generate better music and challenge our creative minds to become even better with what we do as humans do.
Do the search you will find the videos.. I havent heard these precise tunes, but I know its got to be inpainting.. I think what is being done there is they are inpainting a segment then they give the AI enough of a context window (which can be controlled) to give it an idea of what to fill the rest of the song in with.. I just didn't know why until I found your videop.. Will I use it? No, its more fun making new songs with the artists by accident..
Something tells me that Udio will be down tomorrow
This isn't particularly new in the industry or even surprising to anybody, I'm not defending it in any way, but I don't think anybody is going to be shocked into lawsuits tomorrow. It's part of a much bigger issue that's quite pervasive now with so many AI models out there across different mediums
Its NOTillegal. Because its the same as a cover bands rights to recreate music. Its a recreation and not an actual sample from the song.
In theory, yes. But the courts may decide otherwise. Nobody knows what the outcome will be at this point.
Even as a coverband, the venue where you play, will legally have to pay a license to your country's PRO..But this example is tailored to the sync world, in which you can't use copyrighted material without paying a license..
@@RoelandRuijschvanDugteren
It reeally depends which country you are in.
In mine (Italy) you don't pay anything, you can play what you want, *but* a share of the money you'll make will go to the original artist you covered.
If you make zero, you pay zero.
🤡 🤡🤡🤡🤡🤡🤡🤡🤡🤡🤡🤡 just shut the fuck up if you don’t know how AI works
It’s not a cover if you literally have to pull audio data from a song and use the exact sounds from it you nonce
That doesn’t mean that that came off Udio, you could make a song on audio with a voice and then use voice replacement software to replace the voice with John Lennon‘s voice that was trained on something like replay app
Yes I’m sure instead of AI stealing everything like every single AI model ever made this is a case of the AI being framed
But to be fair, if it is a certain individual who is purposefully exploiting the website to generate sound-alikes despite the website explicitly attempts to make it impossible, shouldn't it be the individual to blame?
It just Shows a problem in the system.
For now
It should not be possible to come with the exact same song.
They will fix it.
I still think it is an amazing tool like an assistant for enhance your creativity.
If the artist still develops his creativity and then uses AI to find ideas he wouldn't have think of, like a rhythm , an harmony based on his own piece.
It will be very useful .
“Cleverly Prompted” bro they typed the lyrics in obviously 💀💀💀
If you have used the software yourself this doesn't seem to be the case. If you type in just the lyrics by themselves you won't get those results; there must have been deliberate manipulation of it's programming; also it is questionable why he didn't click into the songs to show the prompting that was used.
In Udio & Suno if you click on a song it will display all the meta data of the song, lyrics, prompts, keywords, etc.
Makes me doubt the legitimacy of the claims in the video.
@@GeorginaJungle yea, something’s fishy
shouldnt you get a new unique song everytime you put in lyrics though? How did the AI make the connection between the lyrics and the song the lyrics belong to, if the AI has not stored the songs in any way. It should just spit out a random new song with the lyrics that were typed, but it doesnt.
@@karlosmartos4646 yes it does create a new song with the same lyrics. I’ve tried it multiple times, that’s why this video is a lil deceiving.
@@karlosmartos4646 You do get a unique song each time. The video does not show the programming/prompting for the song; there appears to have been very specific manipulation to make it go against it's usual programming.
This process of checking generative AI sites for copyrighted works is called “red teaming” and it’s the most fun thing about Ai music 😂
My idea is they are using something like moises to separate the stems and its training it on the stems.. And it using some sort of AI that judges the combined stems to determine if it sounds good, the copyrighted music is being used to train thee AI that judges the quality of the song, and its using the users likes and choices in the music making process to rate the choices, which is key to training the aesthetics of the music..
Finally a video where there is comparison to REAL artists, and not crap or rap artists trying to sound the same and then claming UDIO copies voices. They really "nailed" Beatles, etc. But... I have a feeling UDIO has thought this through somehow, but I dunno... If taken to court they would propably loose straight on. I use UDIO myself (and I am a music producer, of sorts) and I have been able to recreate something similar and even not universally known artist. Need to experiment more.
This is wild. Really disgusting if they trained models with copyrighted songs without paying. But again are we actually shocked if they did? Most of these music tech companies have shown repeatedly that they are not friends of artists.
artists are not friends of artists, AI uses a small percentage of individual copyrighted material for creating ORIGINAL CONTENT, so no matter how many trials occur, AI will end up winning
@@Musikyxwe will see, but there is no right that gives any one to use even „very little” of someone else’s copyrighted work. See how many producers are end up paying 100% and sometimes even more than 100% of their royalties for using even a little bit of sampling.
@@Musikyxhow is this ORIGINAL CONTENT? They literally sampled the Beatles! This is why machine learning is not the same as human learning. You can get the basic idea of something and create something new. This AI model is just pulling from already created things and quilting them together.
@@BrofUJu Let those fat nerds cope, they defended some meme crypto to death they will defend AI copying songs to death, and they will find any hyperbole to justify their fanboyism
I‘d be very interested to see the prompting on those? Because we need to start and fight this training crap! Otherwise we‘ll just be out of jobs!
What about people whose voices are confusingly similar to a celebrity? Are they even allowed to release their own songs? What if I train an AI with a voice that strongly resembles a celebrity ? ( I will never do this, but I like the question)
Anyway: Thank you for your great work!
That's a great question and is exactly what just happened with OpenAI and Scarlett Johansson.
If you can get a non-copyright tune out of the AI then it follows that you could get a copyrighted tune out of it too if all of the training data is treated in the same way. People always seem to find ingenious ways around the attempts to prevent this happening by crafting 'special' prompts. It doesn't help that the creators of the AI don't know exactly how it works.
I might be wrong but this to me shows how powerful this tech actually is. You can create something completely new without trying, and if you try really hard, you can recreate something that's already made almost perfectly. I just don't understand why would people, when faced with such beautiful and promising technology, instead of using it to enhance their creativity, try to manipulate it like this.
Also, this is nothing new. Everyone knows you can recreate some artworks of artists to almost to the T, if you keep trying. But the question is again, why? Obviously, if the tech is powerful, it should be possible to create any style, but wouldn't it be more fun to try and create something completely new?
@@aeorosa this is shit technology that removes all process from creating music. It’s not beautiful it’s just opening the doors to hacks who want to roleplay as musicians.
the rights to 'Last Christmas' are owned by Warner Chappell Music UK , Asked them what they think about it
Is he claiming that the opening two songs were actually produced by Udio?
If so what was the prompt.
@@favesongslist The reason he didn't show how he did it is more than likely he fed udio the full lyrics, then prompted a christmas song from the date of the original, and using keywords to describe a voice like Mariah and probably adding a bunch of keywords to try and point the Ai to Mariah's voice. But I think everyone assumes the Ai was trained on copyrighted material, the big question is if the courts will decide that is illegal or not.
@@favesongslist Interesting that the video creator chose to omit this information: The prompt is most likely just the exact lyrics from those songs, which is practically like instructing the model to create some version of that song. See "Amazing Grace", the #3 Staff Pick on their own website.
@@ooOPizzaHeadOoo Totally agree about the prompt, its what I would do to jailbreak the AI.
Also agreed the issue of pirating copyrighted songs and music. just as most published data like books and video used for AI training.
Use of 'Big' data is such a big issue, the UK government gave big Pharma the right to use personal NHS medical data apparently including names. Also somehow some insurance companies got this data as it was resold.
Like you, where will all this could lead , let alone for the advent of self learning and updating Artificial General Intelligence AGI.
Our governments have not got a clue and are only reactive and then at best slow rather that proactive to the biggest change coming in human history.
@@ooOPizzaHeadOoo OH my detailed reply got deleted :(
Agreed, Jailbreaking is very easy.
The issue is more about where AI is going and the use of 'big' data. let alone the possibility of AGI.
It's not just the copyright issue that's problematic, but also the ethical aspect and the kind of society we want to build. AI shouldn't replace jobs that are genuinely interesting and stimulating for humans. In a logical world, AI should be used to allow people to be replaced in menial jobs, enabling us to focus on creative, philosophical, and scientific tasks. Copyright allows artists to create art in today's capitalist world; if AI renders copyright valueless, humans will stop creating, and machines will take over. It's understandable that musicians are concerned about AI. If AI replaces our creativity, there will be nothing interesting left for us to do in this world, making it a very dull place if AI also takes over other jobs.
People who celebrate AI replacing human art are completely unaware, and this could lead us to a dystopian world. Humans must remain at the center of creation or collaborate with AI, but not be replaced by it. Otherwise, even with new forms of art created by AI, there will be nothing left for us to do. We shouldn't start down this path now. The goal of technological evolution should benefit humans, not businesses like Suno or Udio that are only interested in making money without considering the ethical implications.
It all depends on your perspective, being smarter than other not always equates to being replaced, for example some animals are way faster than humans ,cars are faster too ...did they makes us not wanting to run and compete among a ourselves? AGI if it comes may not really care much about what we do or we don't specially if it becomes a million times smarter it can simply find us boring and just ignore us ,leaving us to our own devices ...people may not care if there is a being smarter than us either ( other than a wounded ego for some) ,just like there is people if certain parts of the world that live in communities without electricity or running water just like hundreds of years ago , many don't care need or want our cars ,cell phone or TVs they are happy with what they have mainly because they are aligned with their own needs and the natural world ...in the same way many people will want to live their life with as little tech as possible or just what they need it ,perhaps machines will make the heavy labor so people indeed will have spare time and will enjoy doing art ,music ,poetry etc "just in their own way and for their own pleasure " ,not really as a way to compete with AGI
you can choose a selection of genres, and if you input lyrics it will create music to fit with those lyrics.
same with vocal styles if you input lyrics and a vocal style it will sing the lyrics in that style
if you teach a band to sing / play instruments on a bands copywrited music they could do all of the learning legally, and liscenced using something like spotify. if you gave this band some copywrited lyrics and asked them to make musing and sing the lyrics in the style they had learned there would be a lot of similarities with the origonal.
how is something like udio learning in the same way any different? it was taught from known music in a similar way and then a user dileberately fed it copywrited lyrics in a dileberate attempt to find something similar to the origonal. its not suprising there are a lot of similarities.
if anything it could be the user at fault for using copywrited lyrics and requesting it be sung in a specific genre/vocal style in an attempt to dileberately reproduce the copywrited song.
these models are taught on music and only store their learnings rather than the music itself so could arguably come under educational purposes and fair use
by the way even if you think the record industry will slap them down, much of it will make its way to open source which makes no money, and can't be removed cause of things like torrent and the fact the music industry as a whole is not significant to fight a world that may fight back
'
It's not a "bug", it's what sets Udio apart from the rest. I personally love the fact that I can have a song generated that sounds just like the artist, but that artist didn't actually perform it. I think you're a wet blanket that's ruining everything for everyone else, and I'm sure I don't stand alone.
Just curious, do you make your living from copyright protected music?
You're not an artist, you'll never understand.
@@Vivaildi I am an artist, and I would be flattered if someone generated a song that sounded just like me. Mimicry is the best form of flattery.
@@prestontedesco9368 yes me too, but not a machine. they're not crafting anything, they are writing prompts and roll the dices. I won't be flattered a machine just found my music in its database and replicate it without any form of appreciation if it is good or not. that's the difference.
I just fed it "A song about Udio getting sued"
Lawyers on deck, court got me waiting
Missed notes in clauses, it's all frustrating
Trying to keep the sound that I've been creating
But these suits got a grip, no negotiating
Courtroom like a stage, the lights too blinding
Finding loopholes, heavy they're reminding
🤣
Our colleague has uploaded several thousand German big band records from the period between 1920 and 1940. When we generate songs like this, we listen to the old records again.
But even when it comes to pop music, nobody dares to pick up a song. You always get the feeling that the song already exists in a similar way.
I m from France. Some Udio french songs voice really looking like an artist called Barbara