AI-Generated Music Is Here. How Are Artists Adapting?

Nastia Voynovskaya [00:02:07] Yeah, absolutely. So Dr. Maya Ackerman is an academic and she was an early innovator in the generative AI space.

Maya Ackerman [00:02:18] So I joined the space commercially in late 2017, but it was only in late 2022 when investors essentially woke up to gen AI being the next hot thing.

Nastia Voynovskaya [00:02:32] She’s a musician and she created this company called Wave AI and they make a software called Lyric Studio that basically gives you writing prompts and helps you build upon lyrical concepts that you’re already working on by giving them prompts and generating next lines and giving them rhyme scheme suggestions.

Maya Ackerman [00:02:51] It kind of invites you to write your own stuff, and whenever you’re stuck, you can get ideas for the next line. It’s not about giving you the answer. It’s about intentionally going to new places.

Nastia Voynovskaya [00:03:01] I interviewed her and she said that about 10 years ago, this was a very niche academic discipline with about 100 researchers experimenting in it, but the trend that she’s seen in the last several years has been that the powerful investors that really have the money to make something happen are throwing their capital behind software that increasingly cuts the out of the creative process.

Maya Ackerman [00:03:35] Starting in 2017, investors didn’t care about it. It’s like, they didn’t think it was real, they didn’t think it would ever gonna catch on. We got to millions of users with Lyric Studios, which is a system that helps people write lyrics. But then once investors realized that it’s real, they wanted to replace musicians.

Nastia Voynovskaya [00:03:52] So the two biggest AI music companies are Suno and Udio, and they both trained their models on all the recorded music that exists on the internet. And music made by both of these softwares has actually charted on Billboard.

Maya Ackerman [00:04:14] And so that’s what our discomfort around it, the sense that it’s here to take over artists, is because that’s where it was designed for, not because that what AI has to be, but that’s why this particular AI was designed to be. So yeah, it’s very unsettling, of course.

Ericka Cruz Guevarra [00:04:31] It’s wild to think about that we sort of went from a time when AI was seen as more of a tool to sort of help artists and musicians and now it’s just doing the work. When did you first start to see AI trickling into the music industry?

Nastia Voynovskaya [00:04:52] I first heard of AI seriously making an impact in the music industry when this completely fictional AI-generated R&B singer named Xania Monet, quote unquote, signed, her creator signed a $3 million record deal.

Xania Monet [00:05:11] Did her best but she can’t teach what it feels like when a father speaks so i took every i love you too

Nastia Voynovskaya [00:05:21] And that was in September, and this was the first AI-generated artist to actually get a recording contract.

Xania Monet [00:05:30] That he never showed me what felt right How was I supposed to set the bar When I ain’t never seen no man fight for my heart

Ericka Cruz Guevarra [00:05:41] There’s a poet behind this AI-generated musician. Her name is Talisha Jones. She contributed to the creation of Xania Monet. How is that different from how humans usually make music?

Nastia Voynovskaya [00:05:58] The creator of Xania Monet, Talisha Jones, just did an interview with CBS.

Talisha Jones [00:06:04] I wanted to reveal myself because I wanted people to know there was a real person behind Xania.

Nastia Voynovskaya [00:06:10] She said, you know, technology is changing. These are just new tools, and people are always resistant to technology.

Talisha Jones [00:06:17] And anytime something… New comes about and it challenges the norm and challenges what we’re used to, you’re going to get strong reactions behind it. And I just feel like AI is the new era.

Nastia Voynovskaya [00:06:29] Talisha Jones says that she writes all the lyrics herself without any assistance from AI. And then she goes into the software called Suno and she puts her lyrics in it and then she just gives it a text prompt. So she will say something like, slow tempo R&B with light guitar and heavy drums. And then so the software spits out. Bunch of options for songs and then she can give additional prompts to kind of help tweak it and then, she arrives at a final product. That’s very different from someone writing the lyrics and then you know singing it and composing a melody and then composing different instrumental parts then you know perhaps editing them on a software like Ableton and arriving at a final product And if you listen to Xania Monet, although the music has resonated with listeners, I can’t say there’s anything original sounding about it.

Kadhja Bonet [00:07:34] Being chosen if he stayed in the hurt was worth holding i called settling a sign of love…

Nastia Voynovskaya [00:07:45] I mean, people have compared her to R&B singers like Keyshia Cole and K. Michelle, and these are people that have spent their entire life training and honing their vocal gifts. What this AI model does is just take that and then learn to reproduce something very similar.

Ericka Cruz Guevarra [00:08:06] I will say it sounds very generic. It doesn’t sound like not a real person, but it sounds really generic. I feel like AI can never do love by Keyshia Cole, you know?

Nastia Voynovskaya [00:08:19] Yeah, or at least not something that would have the impact that it did when it came out.

Ericka Cruz Guevarra [00:08:23] Exactly.

Nastia Voynovskaya [00:08:24] That’s what we think now, but who knows?

Ericka Cruz Guevarra [00:08:47] We’re sort of at a point where AI-generated music is already gaining popularity. What did you hear from the folks that you spoke with about their fears around the economic impact of AI on artists?

Nastia Voynovskaya [00:09:05] I spoke with Joey La Neve DeFrancesco and he is a member of the punk band Downtown Boys and he also is a co-founder of an advocacy group called United Musicians and Allied Workers.

Joey DeFrancesco [00:09:19] Our position is not that technology itself is bad, it’s who owns it, how it’s being implemented.

Nastia Voynovskaya [00:09:25] In his view, it’s the goal of record labels and companies like Spotify to take human artists out of the creative process as much as possible because then they have fewer rights holders to pay for their music. So he really sees this as the end game of the music industry.

Joey DeFrancesco [00:09:44] But yeah, they want an AI artist because they don’t have to pay them, but also the AI artist doesn’t complain. The AI artist can’t unionize, the AI artists can’t do anything.

Nastia Voynovskaya [00:09:57] He talked about how the very origins of the music industry in large part are founded in racism. The music industry going back to 100 years ago has profited tremendously from the creative innovations of black artists that were never properly compensated for their craft. These tools make it very, very easy for companies to further take black sounds and black esthetics and not compensate the people or communities that are driving that creativity.

Joey DeFrancesco [00:10:34] Since there’s been a music industry, the corporations that have profited from musicians have always been trying to devalue musicians’ labor to take the artists out of the picture as much as possible.

Ericka Cruz Guevarra [00:10:51] For Joey DeFrancesco and other critics, huge streaming services like Spotify have a lot to gain from AI. For years, Spotify has padded playlists with AI-generated music, which reduces the royalties it pays out to human artists. And so far, there’s no way for listeners to really distinguish between AI and human-made music. Joey’s advocacy group, United Musicians and Allied Workers, recently collaborated with Congresswoman Rashida Tlaib on the Living Wage for Musicians Act, which would ensure that money only goes to human creators. Spotify says it’s working on responsible AI tools and disclosures for AI-generated music. In a statement, the company said, quote, We want to build this future hand-in-hand with the music industry, guided by clear principles and deep respect for creators. Suno and Udio, two of the most dominant AI music companies, did not respond to KQED’s request for comment.

Ericka Cruz Guevarra [00:12:06] Did you talk to any artists in the Bay Area about concerns around just the heart and the soul of music that just gets lost when it’s AI-generated?

Nastia Voynovskaya [00:12:16] Absolutely, that was a very common concern the artists raised.

Kadhja Bonet [00:12:21] I’m an AI hater. I would put myself like pretty much as far as you can be on the AI hate train.

Nastia Voynovskaya [00:12:28] One person I talked to in particular, Bay Area-raised Toronto-based singer Kadjha Bonet, talked about how when they write music, it’s not just influenced by things they’ve listened to recently or what they want to talk about in that moment. It’s the whole wealth of life experience or the experience of those around them. When I’m writing a song

Kadhja Bonet [00:12:49] I’m influenced by art I’ve consumed. I’m influence by movies and books I’ve read. I’m in influenced by conversations I’ve had recently. I’m I’m, influenced by the walk I had this morning. There’s so much that goes into how I show up to express myself in a day that I think we’re like, we can’t scratch the surface of by putting in like three of your favorite songs and seeing what comes out.

Nastia Voynovskaya [00:13:15] They worry that if lyrics can just be made with the click of a button, then yeah, music will lose its heart. People wanna connect to another human when they listen to music.

Ericka Cruz Guevarra [00:13:25] There’s also like the role of the artist beyond the art. You know, we’ve seen artists play really important roles in social justice movements. Does Kadhja worry about that getting lost with AI artists as well?

Kadhja Bonet [00:13:39] Yeah, absolutely. Like, whoa, what if we can have this Black artist that we know will never stand up for Black rights, right? That we know we’ll never speak up for Palestine.

Nastia Voynovskaya [00:13:49] It’s very, very convenient for record labels and tech companies to have these artists that can just generate profit and that will never speak out about anything controversial.

Kadhja Bonet [00:14:00] For them, it’s incredible. For us as a culture, we lose tremendously.

Nastia Voynovskaya [00:14:05] They and a number of artists have already taken their music off Spotify because of AI, not just AI-generated music, but the fact that the CEO, Daniel Ek, became the chairman of an AI weapons company. Kadhja pointed to emerging smaller platforms. There’s one called Subvert that models itself after Bandcamp, but it’s a collectively owned platform that is democratically governed by artists. And there are other smaller platforms that people are developing. So I know a lot of us have been conditioned to have access to all recorded music at our fingertips on big streaming services like Spotify, but I think there’s a bit more of a splintering now because of these concerns.

Ericka Cruz Guevarra [00:14:58] Well, it seems like, Nastia, whether we like it or not, we are already seeing AI-generated music being rewarded, being celebrated with deals and number one spots on charts. I’m curious if you talked with any artists who are excited about AI and its potential.

Nastia Voynovskaya [00:15:23] I talked to Kaila Love, who’s a hip hop artist from Richmond and a technologist and a tech educator.

Kaila Love [00:15:32] We need to use AI to apply our creativity and build businesses around our music.

Nastia Voynovskaya [00:15:39] She was not excited about AI-generated artists taking the place of human artists, but she does see a lot of potential for AI to serve human musicians by freeing them up from more menial tasks.

Kaila Love [00:15:54] The same things that we’ve had to do as independent artists, which is like book our own shows, plan our own tours, make our own content, find our own network and fan base and be able to contact them directly without the need of a third party.

Nastia Voynovskaya [00:16:08] So she created this company called Goalgetters AI, and it’s more of a marketing services company. So it can help artists generate electronic press kits and create marketing funnels for people to download and purchase their music. So she really sees it as a great tool for the business aspect.

Kaila Love [00:16:31] I’m going direct to consumer, I’m building my own platforms. That’s the way that AI should be used, not to create these clown ass artists. I think it’s just a big distraction on the possibilities of how we can actually use it in a way more meaningful way that creates sustainable futures and upward economic mobility and collaboration.

Ericka Cruz Guevarra [00:16:59] What does this mean for consumers? I mean, do you think people care where this music comes from or do you just think that they care that it sounds good?