Hear me out: Artificial Intelligence may not be bad for music after all

A musical robot displayed at the World Robot Conference held in Beijing, China in September 2021. Li Xin/Xinhua via ZUMA Press

When Barry Manilow wanted to tour in 1982, he was faced with the immense cost of taking an orchestra on the road. So he didn’t. Instead, a larger group of musicians playing traditional instruments was jettisoned and replaced by a couple of guys playing synthesizers. Musicians’ unions hit the roof, saying that synths were costing jobs and the tech should be outlawed. How did that work out?

Around the same time, musicians were exploring the art of sample, the surgical excisions of portions of old songs in the creation of new ones. There was much complaining about that, too. “Creativity is dead!” yelled the naysayers. “We’re doomed to a future of recycled sounds!”

Instead, once the legal issues were worked out (i.e. giving proper credit to those whose works were sampled), a new generation of clever musicians was able to use the new technology to extend the sonic range and texture of popular music.

Story continues below advertisement

Then we run into the introduction of the modern drum machine. Drummers were positive this tech was going to put them out of business. A funny thing happened, though. It didn’t. There were still plenty of gigs for drummers. Meanwhile, musicians started using drum machines for the beats and sounds that humans couldn’t create with their arms and legs. The tech ushered in a whole new set of sounds as well as a new job: drum programmer.

Such Luddite bleats against technological progress are seldom successful. When it comes to music vs. tech, tech almost always wins.

The latest challenge to the status quo is the use of artificial intelligence in the creation of new music. Critics point out the gaps in copyright law. Musicians look at AI-assisted music as scary, cheating, stealing, and even evil. Ice Cube, a rapper who made his bones with generous use of samples, calls AI music “demonic.”

Story continues below advertisement

Yes, there are all kinds of issues surrounding AI and music that have to be addressed. And yes, there are legal and ethical issues that need to be worked out. But now that the tech is here, it’s not going away. It’s time to start looking at the positives AI might bring to music.

Hold on. Hear me out. Here are some scenarios where AI will be very helpful.

Music generation: Human creativity is restricted by experience and ability. No one can possibly draw inspiration from every style and genre of music out there. AI can. With the speech-to-music capabilities of a program like Google’s MusicLM, you can simply say or type in something like “Play me accordion-based industrial dance music with a Lady Gaga-style vocal in Lithuanian” and you’ll get it. I’m not saying it’ll be very good, but by playing around with goofy commands like that, it’s possible that a new type of inspiration will hit.

Songwriting: Everyone runs into writer’s block. Playing with AI could help break out of that rut by presenting ideas for new moods, tones, and approaches. And not just with the 12 notes of our Western musical scale but with lyrical ideas. Sure you can still reach for a thesaurus or a rhyming dictionary, but AI is much faster and much more powerful.

Production Music: There’s a never-ending need for free non-copyrighted music for industrial uses. That includes everything from incidental scores for narrations, corporate videos, podcasts, TikTok and YouTube posts, and even TV and movies. Anonymous musicians labour in studios trying to come up with this important yet innocuous music while trying not to accidentally infringe on someone else’s work. In the past, these libraries of 10-, 15-, 30-, and 60-second music clips were sold to places like radio stations (usually in big binders of CDs and later as digital downloads). A year or so later with all the material exhausted, you’d have to buy/license a new library.

AI will be used to create a virtually unlimited supply of this sort of music. Ask anyone in the production/post-production business how helpful this will be, especially when they’re looking to create material that sounds close to something familiar without actually tipping over in the area of plagiarism.

Story continues below advertisement

Vocal mimicry: Say you’re a songwriter and you need to write a song for Adele. You can compose the whole thing and drop in a reasonable AI facsimile of Adele’s voice to see how the final product might sound. Songwriters and producers will love this.

Licensing voices: Voices will become digital assets to be licensed and purchased. Here’s an example.

I do a lot of voicework and narration and sometimes it gets so busy that I have to turn down jobs. What if I could license my voice, complete with all its intonations and quirks for use in the voicework gigs for which I don’t have the time to do? I can even see a future where my agent just sends out a demo reel to clients who then mimic me for whatever the project. I just cash the cheques as they arrive in the mail. Sounds pretty sweet to me.

The resurrection of dead artists: In 1997, a new TV commercial for Dirt Devil vacuum cleaners appeared, featuring Fred Astaire dancing with a vacuum cleaner. A cute trick, given that Astaire had died a decade earlier. His estate made a tidy sum of money for licensing Fred’s image. That sort of Zelig/Forrest Gump-type technology has been quite popular in film and TV.

AI now gives the estates of Tupac Shakur, Jimi Hendrix, Kurt Cobain, Prine, or any other dead artist the opportunity to license the deceased’s voice to new collaborations that don’t involve finding old, static vocal tracks from the past. Instead, it uses something called “vocal emulation.” Let’s say, for example, Eminem wanted to duet with Pac on a song about the night he was shot and died. Anyone could rap something into the program and it will change the timbre and tonality into something that just sounds like Pac.

Story continues below advertisement

Who would do this? Think of all the thousands of songs that are being bought up by companies like Primary Wave, Hipgnosis, and Concord. This could be part of the way they extend the life of artists and songs.

Musical fan fiction: What if we could get Kurt Cobain to sing Michael Jackson’s Thriller? Think of all the weird parallel universes that we could explore.

These are interesting times, the wild wild west of AI music production. But as we saw with synths, sampling, and drum machines, it will be codified, legalized, and monetized. The future will be interesting.

Alan Cross is a broadcaster with Q107 and 102.1 the Edge and a commentator for Global News.

Subscribe to Alan’s Ongoing History of New Music Podcast now on Apple Podcast or Google Play

Sponsored content