- This topic has 53 replies, 16 voices, and was last updated 2 weeks, 5 days ago by .
-
Posts
-
“I’m on the tail end of writing anyway, cause even without AI, it’s a grind and pays less every year….”
And therein is the ultimate question: If its not your passion, if it’s just a means to a financial end, why do it anyway? What would you write if you didn’t have to worry about how much it paid? Would it be different than what you done to pay the bills?
Spotify and Apple sure do have a lot of “music data” to “share” with AI Learning models. I wonder if they are participants in the party? Their catalogs are substantially larger than Shutterstocks and really, offer more quality.
From a google search about the size of Spotify’s catalog:
“Spotify finished 2022 with more than 100 million tracks in its catalog, according to the company’s annual report filed Thursday (Feb. 2). That’s 18 million more than the 82 million tracks streaming service had the year prior — which averages to about 49,000 new songs per day.Feb 3, 2023”
LOL! With 49,000 tracks being added every day, Do we really now need ai programs adding even more music? It’s just so ridiculous! The last thing this world needs is “more music” and certainly not average AI generated music.
What we do need AI to do is help people who need music, find it quickly through verbal descriptions of what an editor or director would want in their film.
That is a very positive development Art. That’s great that UMG is all over this. I just encourage everyone to write to their senators and PRO’s to raise awareness that music copyrights are under assault by AI leasrning models who are “learning how to create derivatives” of everyone’s music compositions and sound recordings. A “Cease and desist” order against Shutterstock and their AI partners would put an end to this real fast.
And I also agree, I will not stop putting effort into this business until the checks stop coming. As long as checks are getting deposited, I will put in my time.
>I’m on the tail end of writing anyway, cause even without AI, it’s a grind and pays less every year….<
This realization hit me mid pandemic. I found I was working hard for less and less reward. I told myself I would take a break from the studio, which basically turned into semi-retirement (I’m 61). I am satisfied though, having had the chance to create music for income for over 20 years. Good luck when you decide to move on!
>I remember the love and passion I had, as kid, when I first picked up the guitar. That has held throughout all these decades later. <
Absolutely! I started guitar when I was nine years old. Now that I don’t spend all that time in the studio (trying to earn) I’m able to go back to my first love. I’ll play guitar as a passion for as long as I can, not a means to financial ends.
“AI leasrning models who are “learning how to create derivatives” of everyone’s music compositions and sound recordings.”
The irony is that 99.9% of library music (at least RF library music) is already derivative. AI is most likely gathering information in broad style/genre strokes. It would be real stretch to claim that AI copied any individual composer’s work.
The case of “Ghostwriter” creating an AI version of Drake’s voice ventures into the realm of an individual’s “right to publicity,” which falls under trademark law. https://www.inta.org/topics/right-of-publicity/. UMG and Drake’s lawyers should go after “Ghostwriter” and potentially Spotify full-force.
[*** Removed by Moderator ***]
I’ve always been an artist first and foremost, but life and business has made me a stone cold realist as a close second. Had I not been an artist/stone cold realist, I’d be one of the statistical majority who had left this business a decade or two ago due to it being “too difficult” – or I would have picked up a day job to survive.
Reality : Music has been my full time income my entire adult life – AND – I have a family that I’m responsible for. That falls FAR more important than any sense of “artistry” or “passion” in my life. I don’t have the luxury of being an “artist” only. I’m also a businessman, and as such, must take calculated risks, and make difficult decisions based on the market and it’s realities. Most of the artists I know work day jobs to make ends meet.
Reality 2023 : #1 – the ability to earn a livable income in production music sucks compared to 2005. #2 – it cost way more to live in 2023 than 2005.
If AI hits as my colleagues suspect and predict, our artistic passions won’t buy us a cup of Starbucks unless we are well invested and have diversified outside the USD and market. If AI doesn’t overtake us, you will have earned the right to call me a negative naysayer — but even still, I’m still going to be re-directing my musical efforts away from production music into more creative venues, and less in money making venues – cause the money making opportunities are not at a profit level that’s viable anymore (mostly).
Artists, true artists that can ONLY do music still have to eat and pay the mortgage – and that’s me (and several of us). We absolutely MUST deal with realities of the current market and business. [*** Removed by Moderator ***]
Tough thoughts, but that’s where I’m at today in April 2023 as I stare down adding another year to my credentials….
The reality is complaining here just allows one to vent steam and allows for our colleagues to read the complaints and they probably will agree with most of what we all have to say. No sense in ruffling each others feathers as that will not get results.
However, if you do want results (prohibit AI learning models from using our production music to invent their own AI generated music) then write to your senator, and all board members of the PRO’s to raise awareness quickly. AI companies are using our copyrighted music to essentially reproduce new music. this is simply a fact.
Real results is AI companies NOT being allowed to use copyrighted music to teach AI models how to create AI generated music.If you chose to not protest and raise your voice, you are part of the problem.
“AI companies are using our copyrighted music to essentially reproduce new music.”
I just don’t quite get that. Why would companies invest in learning our RF music, when the entire human history of public domain music is available to them?
How would you prove that an AI generated piece of music was based on your music and not a thousand other composers’ works and/or rooted in the basic music theory to which we all have access?
And, because tastes, styles, and the sonic palette evolve so rapidly it seems like a never ending process.
“Wannabees and fakers (sorry, not meant to offend if it does) can work a day job, and pretend to be a composer….even thru the hard times.”
As Art said, that’s a pretty broad stroke. How you earn a living doesn’t necessarily define whether or not you are a composer, e.g. Charles Ives.
“Why would companies invest in learning our RF music, when the entire human history of public domain music is available to them?’
Ask shutterstock , why are they licensing our datasets (our sound recordings, titles, keywords, and descriptions) to AI companies? why are AI companies paying for access to all this data? The answer is quite obvios. Additionally who cares about PD music? The media production community is not screaming for PD music on a daily basis. Sure it gets used from time to time, but the vast majority of demand is for more modern sounding, original production music that support visual media.
“How would you prove that an AI generated piece of music was based on your music and not a thousand other composers’ works and/or rooted in the basic music theory to which we all have access?”
The proof is right there in front of our face. It’s happening before our eyes. shutterstock (owner of p5) is selling our sound recordings and keyword/ title data to AI companies for a figure we do not, nor ever will get access to so AI learning models can learn how to create similar music from our music. Why else do AI companies want the entire shutterstock catalog?
“Additionally who cares about PD music? The media production community is not screaming for PD music on a daily basis.”
You’re kind of missing the point. Think of the entire history of music as an encyclopedia that contains all of the musical knowledge necessary to generate new AI music. What I’m saying is that AI platforms don’t need to ingest vast amounts “McMusic” when they have access to the totality of musical history, most of which is PD.
“AI learning models can learn how to create similar music from our music.”
So can any reasonably competent composition student by simply listening to it.“The proof is right there in front of our face. It’s happening before our eyes. shutterstock (owner of p5) is selling our sound recordings and keyword/ title data to AI companies for a figure we do not, nor ever will get access to so AI learning models can learn how to create similar music from our music. Why else do AI companies want the entire shutterstock catalog?”
What is that proof of, copyright infringement? No. How are you going to prove that a specific piece of AI music infringes on a specific piece of your (registered) music, when AI is creating music based on an amalgamation of thousands if not hundreds of thousands of composers’ music? And remember that a significant portion of that music is already heavily influenced, if not based-on, others’ music. Who gets to sue when an AI platform spits out a corporate track, the composer who ripped off U2 or The Edge? How about a jazz track that starts with an instantly recognizable Tom-Tom beat, the composer who copied it or the estates of Benny Goodman and Gene Krupa (Sing Sing Sing)?
- You must be logged in to reply to this topic.