A couple of years ago, the biggest topic in music was NFTs. They were the future, until they weren’t. Last year, the only thing anyone (including me) could talk about was Web3 technology. This tech is still being developed and rolled out, but we’re not hearing as much about it. In 2023, the obsession is with artificial intelligence. So far, that shows no sign of abating. In fact, we need to talk about some of the latest developments when it comes to AI and music.
Music industry expects AI to be a big influence within the next year
That’s the opinion of Warner Music CEO Robert Kyncl. The Verge has this quote: “Look, you have to embrace the technology because it’s not like you can put technology in a bottle…like the genie is not going back in…Name, image likeness, and voice should have the same protections as copyright and the same simple protections as copyright, but it will take time. I believe it will get there, but it will take time. And in the meantime, we’ll work with the distribution platforms collaboratively to run ahead of that.”
Speaking of which…
Legality of AI-generated voice clones
Earlier this year, a YouTuber named Ghostwriter released a song that used AI-generated reasonable facsimiles of the voices of Drake and The Weeknd. Since then, artists, industry groups, publishers, performing rights organizations, and record labels have been scrambling to sort out the legalities of using AI in music. Musicians are especially worried that someone with a generative AI program will start cloning them without their permission and are looking for legal remedies. What’s more personal than someone stealing your voice?
AI fake voices are a growing issue. Last week, I ran across a new program called YourArtist·AI. This startup has created voice bots based on the actual voices of dozens of real stars including Taylor Swift, The Weeknd, Bruno Mars, and Ariana Grande. You can chat with the bot and it will answer in the voice of your musical idol. And for a fee, you can upload an audio file or a YouTube link and that voice will sing that song for you. Gee, what could possibly go wrong with something like this?
I earn my living with my voice. But now anyone can Google “voice AI clone” and find dozens of programs that will clone me using any available audio of me that’s out there. Is there any legal recourse to this? Not that I’m aware of. At least not yet.
AI music is finding its way to streaming music platforms
After initially pulling some material, Spotify CEO Daniel Ek says the company will not ban AI-made music. He did, however, say that AI shouldn’t be used to impersonate real artists without their consent. But people are going to post those kinds of creations, anyway. Who will be in charge of policing that? Unclear.
Meanwhile, you may have found yourself being served up AI-generated music by Spotify’s algorithm already. Music critic Ted Gioia wrote an excellent piece about how he found himself unwillingly listening to “tinkly generic improvisations” purporting to be jazz. They’re not good, but someone is making money from these things.
I quote (and the emphasis is mine): “This might be funny, except that it’s actually what’s underway in music right now. AI tracks are showing up everywhere, but not because they’re good. There are no devoted fans of this genre. It’s only happening because AI music is cheap, and somebody can make a buck by getting rid of the musicians—provided listeners aren’t paying attention.”
AI music-generating tools are proliferating and easy to use
Want to give it a whirl? There are more and more text-to-music programs available. Just type out a description of what you want to hear and these programs will generate the music in seconds. Humbeatz will turn your voice into a musical instrument. Can’t rap? Try Uberduck,
But wait: What if you could generate music just by thinking about it? If we look at research being conducted at the University of California, it may not be long.
These are exciting tools for creators and have the potential to advance the art of music. But you also know that there are those who are going to use these tools for evil. Because people suck, right?
There's now a music Chart for AI-generated songs
It’s called AI Hits, of course.
It’s not exactly the Billboard Hot 100, but someone is keeping track of what’s out there.
Both the Junos and the Grammys have since declared that songs featuring significant swaths of AI-generated music will not be eligible for nominations. In both cases, any piece of music up for consideration must be created almost entirely by a live human. Fine, but good luck with that. At the speed AI is improving, it may soon be impossible to tell which music comes from a human and which was created by a machine.
AI-generated music is here to stay. So what do we call It?
How about “Syn?” That’s short for “Synergetic” and is being proposed as a brand new genre classification by a company called Music Reactor. I quote: “AI is an advanced tool that artists can use to explore their creative ideas more deeply.” Therefore, the thinking goes, there should be a genre that describes the results of these explorations. Question: Is anyone clamouring to hear artificial intelligence make music?
Outside of those who are experimenting with the technology (check out what real-life human Taryn Southern has done), does anyone really care? You might be surprised.
Last year, Capitol Records signed an “AI-generated rapper” called FN Meka. But that turned out to be a disaster since this creation resulted in a backlash over its stereotyping of Black people and FN Meka was dropped after just 10 days. More recently, though, Warner Music signed Noonoouri, a female AI pop star created by a 43-year-old German graphic designer. The label has since been taking flak that Noonoouri, who is supposed to be in her late teens, actually looks like a sexualized 12-year-old girl. It hasn’t been a good look for Warner.
Japan, however, has been into AI pop idols for a while with two emerging as superstars. Hatsune Miku was created in 2007 and performs concerts as a hologram and has released — wait for it — 100,000 songs, mostly created by her fans. Kizuna AI first appeared in late 2016 but was put on an indefinite hiatus by her creators in 2022.
A little Syn, anyone?
Alan Cross is a broadcaster with Q107 and 102.1 the Edge and a commentator for Global News.