- This topic has 53 replies, 16 voices, and was last updated 1 month, 2 weeks ago by .
-
Posts
-
You are missing the point on no transparency with what Shutterstock is earning from this lump sum, lease arrangement. There once was a time where shutterstock could be checked by a music producer. You could literally license your own track to see if that resulted in a unit sold and a royalty paid on your financials page. Ditto for subscription. We can and do see what titles subscribers downloaded. We do not see what (I assume google, microsoft, meta, and the likes) are paying to Shutterstock as a lump sum contracted fee to “borrow” the entire music library and all data along with each track to teach their emerging AI learning models how to create production music. Shouldn’t these fees collected be published to all contributors on a regular basis? Shouldn’t we know how much p5 witholds under the disguise of E and O insurance? Increasingly, this company can report whatever they please and hide whatever they please in terms of their revenue collected and royalties paid out. All objectivity in the process is eroding more and more each passing year.
“You are missing the point on no transparency with what Shutterstock is earning from this lump sum, lease arrangement.”
Somehow we’ve morphed from an intellectual property argument to a contract argument. On both issues, you need to be able to state a “legally cognizable” claim. So far, the defense would have good shot at prevailing.
This is the exact question that I was asking, as expressed yesterday in Music Business Worldwide.
Thoughts on Ghostwriter, Drake, AI… and why the music industry shouldn’t panic
“The problem of course is that MILLIONS of individual splices of music are already being used to hone AI music modeling.
Unless it’s blindingly obvious or self-disclosed for marketing purposes – as in the case of the Ghostwriter track – how do you prove, as a creator, that your work has been used as source material for AI?”
To refine this, how would you prove, as a creator, the your work was THE source material for any given AI work, unless your work was completely unique and the AI work was substantially similar?
The issue will always boil down to who owns the copyright? I remember when “ACID” came out in 1998 as a loop based music production tool. I did make tracks with this software and I even made some where I literally did not play one single note with my own hands. Rather I just “pasted/ assembled” various parts from various instruments together and tweaked and edited and mixed until there was a “composition”. I have those titles registered as my own “composition” at my PRO. Was I the true “Author”? That remains unclear, but I did make the final assembly and arrangement with 4 bar sections of “parts” pre played by musicians in a studio.
These few tracks still get air play to this day and still generate performance royalties.
So if we all start creating tracks with AI tools and can register the sound recordings as our own copyrights and the PRO’s pay us, everything should be fine. Someone has to curate, mix, and decide what is going to be “prime time” and what will sound like crap. So maybe instead of people “composing” we may just be “assembling and mixing”.
Shutterstock may very well have a goal where they own and are credited as “writer” and “Publisher” of all the AI generated content. Maybe their vision is that they can gradually stop paying contributors. Who knows? There still will be a competition for creativity and usefulness of a sound recording, no matter how it is being created and rendered. It’s kind of like a ‘Painter” opening cans of paint and dumping them on a canvas to create “Art” in 10 seconds. Some people have bought those paintings for $20,000 or $50,000…possibly even more.
People think creating a generic corporate track with a C G Am F chord progression, at 120 to 128 BPM is so simple, yet why is that only a few go one and rack up 20,000 units sold while the vast majority just drift a way with 0 or two sales? It’s because the final mix, the selection of sounds, chord voicings and melodic elements stand out to the listening public and captivate them more so than the others.
So I conclude, with or without AI….there still will be a competition to see who can create the best sound recordings that people will want to buy.
So far, the U.S. Copyright Office has already shot down registration of AI works.
https://www.artnews.com/art-news/news/ai-generator-art-text-us-copyright-policy-1234661683/
But…
https://www.theregister.com/2023/03/16/ai_art_copyright_usco/
“So I conclude, with or without AI….there still will be a competition to see who can create the best sound recordings that people will want to buy.
I agree. So what are you worried about?
I think AI is a going to be a huge problem for music creators, but great business for lawyers….The fact the AI searches the web for it’s content, before it spits out a song or lyrics, is SO discouraging….so, if our lyrics are printed anywhere, etc. Bet it’s just super easy for an entire verse,someone poured their soul into, to be taken & given to someone on chatbot, etc.for their AI song… when we see MS techs & G techs like a deer in the headlights at an interview…time to worry!!
Although I have some general concern that AI may have the potential to disrupt the production music world, I’m FAR, FAR more concerned about much larger risks that LLMs pose.
The guys behind the documentary the Social Dilemma recently did a seminar (called ‘The AI Dilemma’) about the very real risks LLMs post to society at large. The outcomes they describe combined with the real world examples they provide have far scarier ramifications than simply being put out of work by an AI designed to write music. Keep in mind that numerous high profile people leading some of tech have been urging for a ‘pause on AI development’ as of the end of March. And, one of the two presenters testified before Congress only 3 weeks ago.
I also realize this post probably has a tone of someone freaking out and screaming that ‘the sky is falling’… I’d encourage anyone concerned, but still skeptical to watch the seminar (below)… I was previously in that camp myself until watching it, I personally found it deeply disturbing.
The AI Dilemma:
Tristan Harris’ recent congressional testimony:
@soundspot – thanks for that video. It is pretty much spot on from what my AI programming friends tell me. And scary as $#@!
Perhaps more focused to this thread (but nowhere near as important as your video) :
Hilarious, sad and telling of where we are ultimately headed.
Wow – that’s a pretty powerful quote. I think that’s one that’s gonna stick w me.
Well it looks like that dataset payment that we got from Pond5 was just a one time deal. No payments this month so I guess that was that.
Hey at least we can all move on from composers blaming other composers for the current state of the production and royalty free music industries
https://youtu.be/d7HblZlVgyI latest from the Pond5 dataset apparently.
- You must be logged in to reply to this topic.