Making music with AI is fun, but it also brings some worries

AI News


SAN FRANCISCO — Fun fact: The closest thing this newspaper has to a theme song is a John Philip Sousa march that you've no doubt heard. It's a classic, but there might be something better.

Unfortunately, I'm not a songwriter, so I turned to AI.

This week, Suno, an artificial intelligence startup that lets you create songs by just entering a bit of starter text, released the iOS version of its app, making it easier than ever for regular people like you and me to create music on the fly.

That's probably not welcome news to the several record companies that sued Suno in late June, arguing that the company's tool can generate songs because it learned how to do so by studying the countless songs that they own. (Suno, for its part, calls its technology “innovative.”) Still, the app is still available to download for free for now.

And since the app was released a few days ago, what started as a silly experiment in making catchy journalism-themed songs has turned into a bit of an obsession for me. After all, using an AI to make full-fledged songs on impulse is really fun, but it's also starting to change my relationship with music in ways that I'm not too happy about.

I'll explain what Suno can do and why I felt a little uneasy living with it.

Getting started with Suno is easy: just create an account, decide if you want to pay extra to create more songs every day, and fill out the 200-character prompt.

Generating these songs can take anywhere from a few seconds to a few minutes, depending on whether you've paid for a higher level of service. When you request one, two tracks are always generated and available for you to review.

Your musical tastes are probably different than mine, but I already knew what The Washington Post's first attempt at a new theme song was going to sound like: bright, jangly guitar was a must, along with a meandering, adventurous bass line and journalistic lyrics.

But when they asked Suno to create exactly that, they produced two generic pop-funk tracks that used the words “bright and jangly” as lyrics rather than instructions.

Get caught up in

Stories to keep you up to date

[Listen for yourself: Washington Funk 1, and Washington Funk 2.]

Maybe the genre wasn't the right one. I then typed this prompt into Suno to see if it would copy a specific artist: “Early 2000s Paramore-style pop punk, high energy, female vocals, lyrics about The Washington Post.”

Neither song immediately felt like a Paramore knockoff, perhaps because Suno completely ignored the desire for a female vocalist, but the songs still felt like something you'd hear in high school and featured surprisingly catchy choruses.

Telling stories you need to know

From the city to the world and back to the city

There is no turning back on that page.”

[Listen for yourself: Postamore 1, and Postamore 2]

I wanted to keep those lyrics (and a few tweaks) as a final attempt, so I opened Suno's “Custom” mode, pasted them in, and tried again. (Interestingly, if you want Suno to base a song on a complete set of lyrics, the website instructs you to be careful to only use AI-generated lyrics, but the app says nothing about that.)

Now for the rest of the instructions. I felt I should go a step further, so I requested that the style of music include the following elements: “J-Pop, Math Rock, Female singer, Anime theme, Instrumental intro, Guitar solo outro.”

And for the first time, the results on Suno felt like a perfect embodiment of what I’d given them in my prompts, except when both tracks suddenly ended, went quiet for a moment, and the fake guitars started up again for one last rehearsal.

[Listen for yourself: Washington! Post!! OP1, and Washington! Post!! OP2]

Well, certainly, none of these will ever replace The Washington Post March, but if any of them have a chance, it's Postamore 2.

After I finished writing songs for AI Journalism, I found myself messing around with Suno, making silly songs with nonsense lyrics and trying to recreate the style of some of the one-off tracks I liked.

But I soon began to feel like I was using the AI ​​too much, and sharing the results too much. My wife was having a rough day, so I sent her a lovey-dovey AI song that included our goofy nicknames to cheer her up. I made up some really bad rap lyrics and sent four Suno songs in a row based on them to my friends.

Then it occurred to me: I could easily see myself writing songs and sending them to people as thoughtlessly as I send out a barrage of emojis.

Music is a force for good, for joy, healing, activity and reflection. Is this generation of sloppy music somehow devaluing music in my life?

Max Vehni, one half of the indie-pop duo Slenderbodies, talked me out of that idea.

“Music is a way for people to express themselves,” he says, “and if it's another way for me to communicate with my wife, I think that's really great.”

Clearly, Vehuni is no AI music pessimist: He's tried out Suno and similar services for personal projects, and says he sees great potential for AI as a tool to enhance artists' songwriting and production.

And while Suno has been sued for using copyrighted music as training data, the company was quick to acknowledge that its process is not all that different from what a human would do.

“Artists draw the line: 'It's OK for an artist to be influenced by me, or for a human to be influenced by me, but when a computer is influenced by me, that's not OK,'” he said. “Is this something to be for or against? I don't know.”

But that doesn't mean I don't have other worries. For example, my remaining anxiety stems from worrying that I'm annoying artists I love by making music that sounds like their music but isn't.

Fortunately, Slender Bodies makes most of its money from touring, and the band is fortunate to have a fan base that's willing to support them in the “post-AI music apocalypse,” Vehni said.

In other words, choosing to directly support the artists you care about is more important than ever.

Still, he worries that record companies could sell their copyrighted song catalogs to AI companies and in return gain access to a model that allows them to create synthetic music without paying royalties. Or that streaming services could create and promote their own synthetic artists and pocket the revenue (he's not the only one to wonder about this).

How this all plays out remains to be seen. Either way, big tech companies, the music industry, and the rest of us will have to continue to grapple with the AI ​​music creeping into our lives.

“We took it out of the box, and I don't think we're ever going to put it back,” Behni said.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *