As AI music generators get more sophisticated, the issue of copyright ownership is becoming increasingly complex. With AI tools becoming capable of creating entire songs without human input, questions arise about who owns the rights to these creations and how they should be licensed and monetized.
AI-Generated Music
Music created by artificial intelligence (AI for short) is becoming a big topic of conversation. Perform an internet search for AI music generators, and you will see many websites that create AI music at the touch of a button.
As a musician, many ethical and legal complications arise when contemplating the use of AI music. The first question being – should you? Really! You are a creative individual. Why would you let a machine take away your creativity and individuality?
You are only fooling yourself if you think music produced by a machine is yours, even if the website providing the music is telling you otherwise. AI music will not reflect your personality – which is the essence of artistic expression.
Creativity and Copyright
Let us talk about AI music creativity and copyright (the latter being a legal term for protecting a unique creation in the eyes of the law).
For every artist, surely the goal is to impose your ideas on the world to let the masses know what you are about and how unique you are as an individual.
If you write a song, then you know its origins. You felt the emotion upon its delivery. You agonized over the words, the melodies, the chord progressions, and the stylistic voicing to find the feeling you wanted to impart. You are assured beyond any doubt that you created the song.
The musical work never existed before you brought it into the world – how cool is that!
By contrast, AI music will not give you that same satisfaction. An AI piece of music has been concocted from many existing songs and is basically one big ‘mash-up.’
An AI cannot create something from nothing. It must be fed pre-written songs to learn from and use as a basis for new work. This means you cannot know for sure that AI music is unique enough to satisfy the lawmakers, a doubt that will always lurk in the back of your mind.
With every passing day, there is a chance you could receive a call or an email saying, “Hello, you have broken the law by copying someone else’s song.” If you do get that call, you will have no argument against the judgment because you will not know where the AI music originated from.
You can never be sure you did not copy the said works unwillingly. You will have no defense because no paper trail proof exists. The work was created remotely by an AI analyzing existing songs.
Furthermore, you probably do not own the copyright to your AI song and may never do. As of the time of writing, ownership of copyright to AI-generated music is still to be established within the law in most countries.
Do not pay for Copyright – at least not yet.
To look further into the subject of AI music copyright, there is a disturbing trend for AI music websites to ask for money to transfer copyright ownership to the end user.
Before paying them a dime, it is vital to reiterate that currently, the legal world has not recognized that AI musical works can indeed be copyrighted!
Think twice before paying for ownership of AI music when ownership has not yet been established.
Equally as important is to know that at any time in the future, you could be sued for copyright infringement of existing songs. The legal system still has to establish whether copyright laws have been broken in the generation of AI music, as they are based on content from established and protected works.
You are being asked to pay for a copyright that does not currently exist, one for which you may be liable and sued for infringement of intellectual property if the music you use is too similar to an existing song!
Creative Pressure
In the musical world, independent musicians are encouraged to replicate successful artists to earn money and find success for themselves. No more so than in the world of synchronization.
Sync, as it is known for short, is the act of placing music to images – think TV adverts or film scores. If you have ever tried to write a piece of music for a sync opportunity, you will know that it is common practice to be given a reference track as a guide.
Basically, the production company wants you to recreate the said track without paying huge royalties to the original creator; and we more commonly accept this to be okay, which is why AI music generation is so tempting. It is becoming the norm to use other musicians’ songs as the backbone for new work, and the more it happens, the less we are shocked by the practice.
We are becoming conditioned to hearing similar songs. So, AI music may easily creep into our daily lives if we let it.
We like exciting new things and want them to land in our lap without much effort. If a machine can perform a task in seconds that would typically take us many hours, and the results are exciting to our minds and ears, then we are all in.
Again, do not be deceived. AI music is not yours.
It is not your personality that is reflected – just a mash-up of previously recorded successful songs. You are selling yourself short if you pass AI music as your own.
Where AI Music Works
However, I see various legitimate uses for AI music creation in the studio.
If you create video content that needs a backing track, it makes sense to grab a sound bed instantly with no production costs. You do not need to own the copyright, and the AI music sounds great as a backdrop to your images.
This makes economic sense, but the guy who currently gets paid for writing your music will be losing out, which leaves you with an important ethical decision to make.
As a songwriting tool, too, AI music can be useful.
If you already know your unique voice and want to explore new avenues, then yes, generate an AI musical idea in your genre and think of it as a starting point. View the work as a collaboration. Remold the sounds and ideas to how you would normally write. Take it, run with it, and make the voice your own.
The new song should sound, feel and fit into your back catalog and future content as if it came from your mind totally; it should be your personality that is dominant in the works.
Try your best to be an authentic artist. Believe in yourself. Believe in your talent. Speak musically in your own voice and make your mark on the waiting world. Create songs you can be proud of, songs that you know came from your heart and mind.
Know beyond any doubt that a song represents your vision and yours alone, one that can never be taken away from you, then sit back and glow in the resultant satisfaction.
AI Music Copyright – FAQ
Do I own my AI-generated music?
The issue of AI-generated music still isn’t covered by copyright laws. Think twice before using AI music in your projects or paying someone for AI music copyright.
Generally speaking, AI music was not created by a human factor, so it can, at least in theory, be considered copyright-free. However, this is still uncharted territory, so our advice is to be very careful when using AI-generated music in your work.
Can you copyright AI-generated music?
The debate surrounding AI music and copyright laws is currently ongoing. While some argue that a license is required because the AI’s output is based on preexisting musical works, others maintain that using such data falls under the “fair use” exception in copyright law.
However, advocates for AI claim that the resulting works are transformative, meaning they do not create substantially similar works and have no impact on the market for the original musical work.
Our advice is to wait until AI-generated music is covered by copyright laws before using it in your professional work.
If you are interested in audio production, please check out my books, Audio Mastering in a Project Studio: A Practical Approach for a Professional Sound and Template Mixing and Mastering.
33 Comments
Jed
onAll AI based products do not have any Owners copyrights (Master copyrights) Even the AI Arts, as they are till now a subject for ETHICAL and Juridical development/proceedings.
After the ending of those processes, you can simply Lose all the rights that you own on an AI product after taking its model (Generators) down (Including all their past sold derivative)
I Also do not advice to use AI generators to make arts for your albums and releases. They can simply face the same destany.
ASJ
onJust curious, how possibly can somebody tell a picture was AI generated? Yeah, I know about realistic human hands and faces, but if you generate a picture of a table and chair, could you ever guess it was artificially made?
simon taylor
onYes, it is difficult to tell the difference already and AI is only in its infancy – imagine how hard it will be in years to come!
But the issue here is ‘ownership’. A piece of art (musical or otherwise) needs proven ownership to make money from as an artist. If you cannot prove ownership you cannot make income from your works. You could spend many hours and lots of money creating and promoting a product without being aware that you do not own the master rights. Many musicians are not aware of this, especially when AI music generating companies are transferring copyright ownership for a fee (a copyright they do not own!). You would naturally assume they own the copyright and are assigning it to you – not true as the law currently stands.
MultipliedCow
onUsing AI to write your song is just like an “artist” walking into a studio and having a song handed to them. They sing the song and then it’s “theirs” and they bank off of someone else’s hard work. It’s basically karaoke. I compare that to a coloring book. The art is done, you’re just filling it in and claiming it as your own. If you feel good about having someone else write your music, AI should be the best thing ever. Developing music from scratch is more impressive and respectable.
ASJ
onSo, Adele is a fake artist doing karaoke because she sings a song written by somebody else? And any classical pianist performing a music sheet written by Mozart is colouring a book? I believe your argument may be flawed. Putting aside AI generated content, claiming that you are less respectable because you don’t write the music you perform is questionable, to say the least
simon taylor
onIf an artist sings a cover version or a song written by another creator, that is fine, everyone involved knows who the creator is and can reward them for their unique creativity. This is the difference. With AI it is copying not creating and there is a big unknown regarding ownership!
Totally Not A Robot
onWhile I totally agree that using a.i. tools to generate entire pieces shouldn’t be thought of as a genuine means to produce your own music, I question the rationale of “creativity” when these discussions come up.
You say “AI cannot create something from nothing. It must be fed pre-written songs to learn from and use as a basis for new work.” but that’s the same way humans do it. We don’t generate artistic works from the ether, we do it from the music we listen to, the art we consume, how we were raised, the political climate, the sounds of nature, etc. It’s not far off from that process of machine learning. Computers are a human invention meant to emulate and more efficiently process things we already knew how to do.
I would argue a.i. as a genuine extension of ones own creativity is more than valid, and potentially even a collaborative effort.
However, as with most things, capitalism and the crap that comes with it, like copyright, is making people hostile towards innovation and change as it always does.
simon taylor
onThere is truth in what you say. I would only add that humans have been making music for generations and we know the boundaries and the difference between copying someone’s work and making music our own – we accept that we are influenced by our peers and hero’s, but try to give life and our unique personality to a work that is our own. AI’s don’t have that in-built sense. And yes, it is true that each individuals views on how music should be delivered to the listening world will influence how AI generated music is used and received.
Kat
onThe problem is, I don’t, really. I don’t know the difference. When I am writing music, I am often worried that it is derivative–not only that it is cliche, but that some part of it is really too close to a part of an already existing piece, perhaps one I know. I can avoid complete plagiarism, but not this type. The way copyright seems to work in the modern music world does not help matters, even if it really isn’t the proper metric of plagiarism anyway.
And I always, always, give the music my own personality; I’m not enough of a polished professional to do otherwise, even if I wanted to (and too much a diva to want to)!
Dr Gray
onI think the difference is that when a human creates an original peice of music they are building on what has gone before, but when they are truly original they have innovated and pushed forward the boundaires of the entire body of written music. Mozart was influenced by what came before, but he pushed boundaries in the way that an AI coud not. Ditto for the Beatles or Wire or any other innovative artist you can think of.
Of course not every peice written by a human bieng pushes boundaries and since time immemorial there have been writers who have produced little more than “a mash up of exisitng ideas”, but they have largely faded and it’s only the true innovators who remain. I think we are some way from AI being innovattive, being random and coming up with something original but undirected isn’t onnovation. When one of the blindfold monkeys typing randimly produces the ccomplete works of shakespear that ia not creativity…
CMD
onIf by using Royalty-Free samples that you have paid, you won’t never own 100% that part of your production you used these for, imagine the Copyright rabbit hole AI music will have to transit.
When buying instruments from NI, Spitfire, UVI, etc. you have ONLY been allowed to use these.
There’s a “stem” in your production that have been recorded by them, they have the Copyright of the sound recordings (one-shots, multi-sampled instruments or loops) they will always be legally protected to demand any % of your earnings if you had some success.
Not to mention loops, as these have 1. the intellectual property add on of being written music 2. WhoSampled, ironically the best source for loops producers to hunt a holiday bonus from someone.
Soon there will be some WhoGenerateIt or such, and things will just get dirtier.
simon taylor
onOf course, nothing should get in the way of creating great music – use whatever gets you there. There are parallels between AI now and early hip hop recordings/artists who used samples knowing that they did not own the copyright – they didn’t care, just wanted to make great music and show their personality and get their artform liberated. But fast forward 40 years and they do now care because they put all the hard work into building a following and making money only to end up in court and be sued for copying (sampling) other works. So now most artists sampling will declare the samples used and pay the creator a percentage – which is where I see AI music going, everyone involved in the process will get a slice – and the musician will get a tiny crumb that is left – sound familiar!
Allen
onA very interesting and informative article/ debate. About a week ago I came across a YouTube video, which advocated downloading/coping music from any website and using it with AI to create videos. I felt compelled to place a comment, to warn anyone viewing the video that copyright laws existed. Both in the music and use of video material. I am not a musician or videograher but I have a very long-standing love of all art-forms and a relatively good knowledge of copyrights in general.
How long will it be before a synthesiser preset bank digital or analogue becomes used by AI and the “synthesiser patch” is sited in a copyright claim? After all a sound designer created that sound.
AI is fascinating but is also dangerous, any means to get the job done, is far from true art,
ASJ
onIn electronic music, artists have been evading copyright for decades by sampling stuff and “flipping” the sample, by repitching, slicing, rearranging and creatively processing content not originally ideated by them. Chicago house, old school Hip hop, Jungle/Drum n Bass and many other styles were born out of this practice. Anything from Daft Punk prior to “Random Access Memory” album featured sampling from disco music tracks. With AI, one could generate samples and, by flipping them, obtain something new that can’t be framed for copyright infringment. Obviously for other genres this could not be always possible (not really suitable for a soft acoustic piano piece, for example).
Tomislav Zlatic
onGreat point! I also think that AI music tools could be a great sound source for sampling an mangling.
Simon Taylor
onYes, agree, there are many positives to seek out. As a learning tool too. Some of the AI websites will break the music down into individual parts and allow the exporting of MIDI tracks too. This is a great way to learn how songs are built track by track and how to arrange and organise music so there is always something going on – a great way to analyse music. Try exporting the AI song in MIDI and assigning your own sounds, learning from the arrangement and breaking it down, chopping the sounds and resampling to make it your own!
ASJ
onThis is very cool indeed
Simon Taylor
onYes, so cool, the first thing that jumped out at me when trying AI music as an experiment was how great the arrangements were. They were always top class. If the sounds weren’t great, the arrangements always were. Getting a quality arrangement from the start is key to making the subsequent mix and master easy. The best arrangements will mix themselves. Without any instruments clashing the balancing is straightforward. A poor arrangement and you’ve got to work hard to force the magic to happen. I have to say I am neither for nor against AI music generation – let’s see what it brings, but for me, if I was starting out right now, I’d study how AI arranges musical elements.
David
onI agree that using AI to compose music removes the human element regardless of how creative the prompts may be. Regarding AI music being “one big mashup”. Every song that any human being writes could be considered one big mashup.
All of us are influenced by every piece of music we have heard. It’s no different than training an AI. When we write music it does not come from nothing. It is informed by every piece of music we have ever heard.
Simon Taylor
onjust my opinion, but as a human I believe we want to be honest and unique and create music that is personal to us. I think we musicians will create music in an ethical way for ourselves. Agree, we cannot ‘unhear’ everything we have in our lives, but we always strive for something great and unique. My worry is that whereas we musicians create for the love of music. What do AI websites create for the love of? I suspect in the end it will be for profit alone. Unlike humans an AI can ‘unlearn’ and ‘reset’ to focus on one new thing. I can see it going where one AI is just fed songs from one artist – say Prince, for example, and you can buy your own Prince song to sing over that is indistinguishable from a song written and performed by Prince. A good or bad thing!! Depends on your take.
Glenn Stanton
onpretty sure if you are using Google or Microsoft AI tools, then they own it. and they have the small print you probably didn’t read as well as the lawyers to back it up.
Simon John Taylor
onYes, just to be clear we are talking about music creation AI. The AI’s in question have been fed existing hit songs already written and protected by copyright law without the consent of the copyright owners – not legal. This is problem one. The second problem is an artistic work is created as a consequence of using musical AI – problem number two, who owns the works? The guy who wrote the code for the AI, the AI itself, the musician who pressed the button? The hosting website? In current law (depending on country) AI artistic output isn’t copyrightable because the answers to these questions haven’t been answered.
Junkyard
onI’m confused why you think I “do not need to own the copyright” if I need a backing track for video content that I’m creating.
Simon John Taylor
onYes, just to clarify – you would need to own the copyright for a backing track for video if you wanted to monitise the audio from the video. I am thinking of vlogger’s, YouTube, TikTok etc where the video and backing track are not related i.e. the creator just wants a soundbed to their images and are not concerned with moitising. If you are making a video to your song as a musical artist then yes, you want to own the copyright (or the company you have licensed it to).
Ivan Mattheus
onThis is a great subject to deal with, but if we think more about the question of a song lyric, let’s assume that I have a great idea, or a ready-made lyric that needs adjustments and things like that and ask ChatGPT for me help complete or settle these things, would that be wrong? Would I be breaking someone’s rights because of a word or a small verse the same or similar to someone else’s?
And one more point is, if I followed this idea, I would be missing the feeling of the music, the experience, the existence, the essence, after all I didn’t write it completely.
(I apologize if I have problems, I used the translator)
Simon John Taylor
onJust my opinion, but I feel it is about degrees of change/similarities. To change a few words in a lyric using ChatGPT would be no different to using a rhyming dictionary or app to help finish a song. The works would still feel like your creation. To ask ChatGPT to write you a verse, chorus or complete song could result in you losing connection with the song. That would depend on how you interpreted the lyric though, and if you are using a whole GPT chorus, you could never be certain that the chorus has not been taken from an existing work.
Zsolt
onCopyright “laws” were an ugly mess held up by precedents, conventions, and sheer power even before digital sharing and AI existed. Digital copying and sharing art combined with AI merely stirs up mess to the degree that its ugliness becomes apparent for all of us to see. There is hope in all of this you know. For ages, every time the system of music creation and consumption got reworked, we have always found a way to work it into something new that serves the interest of the few over the interest of the musician or the music lover. Hopefully, soon we’ll have another chance to create something better that serves the humble music lover and not the money/fame/power lover. Based on the past I do not expect a lot, but hey, there’s another opportunity brought us by technology breaking the old toys and bringing new ones.
Simon John Taylor
onGreat comment and so true. If you look back through musical history, most musical trends were driven by new technology. AI could be the new catalyst for something great. It would be wonderful if this transpires to inspire and elevate music for musical ears and not just server to generate more ‘content’ more quickly. Wouldn’t it be great to get music back to valued music and not simply content to be generated.
foxAsteria
onBahh maybe stop trying to own everything. Not even your firmly held beliefs belong to you, let alone the music “you” produce by regurgitating everything you ever consumed.
As long as we keep deluding ourselves into thinking money has more value than time, we will keep missing the point of art and artistic process.
Simon Taylor
onGreat comment. Yes agree, there is nothing more valuable than time. Play the game however makes you happy. Create art however ‘feels’ right. The most important thing is to enjoy your own ride! There is a great freedom in creating and letting go without any further thought for what happens next. It focuses the mind…create/release, create/release…create/release……..
David Harper
onOoh this is going to get crazy messy. Copyright infringement has been all over the map in recent years. People have lost in court simply for grabbing a vibe that belonged to an artist. Melody note counting will be easier than ever with new technology. When big money is on the line, we’re going to see big lawsuits. So far the Library of Congress refuses to grant copyrights to anything other than human composition. It sure will be interesting to see where all this is going. I don’t find myself hopeful in terms of songwriter growth. Listeners’ expectations have already been lowered below absurdity. AI arriving now, on the heels of streaming just completes the mediocre puzzle. The great ones will rise to the top. The rest of AI players will be no more valuable than clip-art.
Simon Taylor
on…and how mind blowing the amount of AI music that can and will be created. As a quick test I just generated a piece of AI music – it took 12 seconds!!! 12 seconds!!! Think of all the users, all over the world, clicking buttons every 12 seconds!! So no less numerous than clip art, agreed, but the difference I see in the analogy is that AI music is perfectly usable in the right context and difficult differentiate from bespoke written tunes – depending on the source site. AI music can be of a excellent quality and whether it is profitable to the individual songwriter will come down to chance and capturing imaginations as it is now.
Krisna
onHello, I want ask, Can AI cover song get monetize on youtube?
How about I wrote the song, then record song, then I generate my Acapella by AI cover for examlpe Ariana Grande model, this is legal? Can monetize ?