Deeply provocative post. I was struck by the thought that Spotify could adopt an Amazon model fairly easily — viz., use their own analytics to track which songs by real artists are at any moment attracting streams and have generators produce simulacrums that enter the algorithms or playlists with a little extra heat. Since these generated songs would be "owned" by Spotify, the company wouldn't have to pay out for those streams that would, invariably, eat away at listens that might otherwise accrue to real musicians.
This move would be utterly evil, but they have little incentive NOT to be evil. The major labels might even be on board with such a practice, as it might increase the perceived value of their catalogs full of music from real human beings.
The RIAA has already indicated they see the output of generators that have been trained on major label music as infringement, so that specific application of the tech will be tough to pull off in the near term. I have to check into the specific laws a bit more to make sure, but I think that's why we aren't currently seeing Spotify investigate tools that could seamlessly transition two songs together the way a DJ would in a club. The primary purpose of the major labels/rights holders at this point is to protect the value of their existing catalogs by litigating as aggressively as possible. If Spotify themselves were to get in the business of infringement, the majors would likely just pull their catalogs from the service in protest.
Spotify is already behaving exactly as you describe with respect to podcasts and audiobooks, though. An audiobook company that Spotify bought in 2021 recently allowed Apple to use their huge cache of audiobook recordings to train a new "AI Narrator" product that can automatically generate audiobooks. I'm sure they'd love to be doing the same thing with music, it might just require a circuitous path because of how intensely regulated major label music is online. Maybe independent operators will have to overwhelm the platform with generative product that was illegally trained on copyrighted music in order to get the ball rolling, maybe Apple or Google or Amazon will simply just buy one of the majors in order to use their catalog as a dataset. I'm sure we'll find out soon!
Great article. As one of the independent artists receiving less than 1000 streams a month, I can add a tiny bit of insight. I know a number of songwriters and musicians, some quite talented, and for them to achieve even a modest number of spins, they have to hire a promoter to get their songs playlisted. It’s pay to play, roughly $1000 per song, and there is little hope of a return on investment. And those plays may well be passive and not generate a fan of your original music. The flood of 100,000 daily uploads comes mostly from people who are paying distributors like Distrokid and Tunecore to have their songs heard. So the market has flipped, and it is the musicians who pay the listeners to hear their recordings. This was already in motion before the recent advent of AI generative music, because of the sheer volume permitted by digital streaming distribution for 20 bucks a year. However, excepting the devastating impacts of the pandemic on touring, live music survives. I can still get paying gigs and AI hasn’t taken that away. Yet. Last point, it seems that all artistic endeavors that can be represented and distributed electronically are being devalued by AI generative substitutes. Music, visual art, photographs, film (my cousin is a Hollywood screenwriter, currently on strike with major concerns re the impact of artificial intelligence on story and script generation, as well as directing and acting). Even weddings and parties, cornerstones of human interaction, can be experienced virtually, without the warmth of a true embrace. My hope is that consumers will, like JP Sousa, see the value and creativity within real in-person experience, and see the generative media for what it is. A tool.
Buen artículo. Puede configurar el tono de llamada de su teléfono desde las divertidas melodías de tono de llamada aquí: https://descargartonosdellamada.com/
Deeply provocative post. I was struck by the thought that Spotify could adopt an Amazon model fairly easily — viz., use their own analytics to track which songs by real artists are at any moment attracting streams and have generators produce simulacrums that enter the algorithms or playlists with a little extra heat. Since these generated songs would be "owned" by Spotify, the company wouldn't have to pay out for those streams that would, invariably, eat away at listens that might otherwise accrue to real musicians.
This move would be utterly evil, but they have little incentive NOT to be evil. The major labels might even be on board with such a practice, as it might increase the perceived value of their catalogs full of music from real human beings.
The RIAA has already indicated they see the output of generators that have been trained on major label music as infringement, so that specific application of the tech will be tough to pull off in the near term. I have to check into the specific laws a bit more to make sure, but I think that's why we aren't currently seeing Spotify investigate tools that could seamlessly transition two songs together the way a DJ would in a club. The primary purpose of the major labels/rights holders at this point is to protect the value of their existing catalogs by litigating as aggressively as possible. If Spotify themselves were to get in the business of infringement, the majors would likely just pull their catalogs from the service in protest.
Spotify is already behaving exactly as you describe with respect to podcasts and audiobooks, though. An audiobook company that Spotify bought in 2021 recently allowed Apple to use their huge cache of audiobook recordings to train a new "AI Narrator" product that can automatically generate audiobooks. I'm sure they'd love to be doing the same thing with music, it might just require a circuitous path because of how intensely regulated major label music is online. Maybe independent operators will have to overwhelm the platform with generative product that was illegally trained on copyrighted music in order to get the ball rolling, maybe Apple or Google or Amazon will simply just buy one of the majors in order to use their catalog as a dataset. I'm sure we'll find out soon!
Great article. As one of the independent artists receiving less than 1000 streams a month, I can add a tiny bit of insight. I know a number of songwriters and musicians, some quite talented, and for them to achieve even a modest number of spins, they have to hire a promoter to get their songs playlisted. It’s pay to play, roughly $1000 per song, and there is little hope of a return on investment. And those plays may well be passive and not generate a fan of your original music. The flood of 100,000 daily uploads comes mostly from people who are paying distributors like Distrokid and Tunecore to have their songs heard. So the market has flipped, and it is the musicians who pay the listeners to hear their recordings. This was already in motion before the recent advent of AI generative music, because of the sheer volume permitted by digital streaming distribution for 20 bucks a year. However, excepting the devastating impacts of the pandemic on touring, live music survives. I can still get paying gigs and AI hasn’t taken that away. Yet. Last point, it seems that all artistic endeavors that can be represented and distributed electronically are being devalued by AI generative substitutes. Music, visual art, photographs, film (my cousin is a Hollywood screenwriter, currently on strike with major concerns re the impact of artificial intelligence on story and script generation, as well as directing and acting). Even weddings and parties, cornerstones of human interaction, can be experienced virtually, without the warmth of a true embrace. My hope is that consumers will, like JP Sousa, see the value and creativity within real in-person experience, and see the generative media for what it is. A tool.
Extremely well written, the ending in particular fucked me up. I'm going to ask my mom what music she likes.
Buen artículo. Puede configurar el tono de llamada de su teléfono desde las divertidas melodías de tono de llamada aquí: https://descargartonosdellamada.com/