The episode in which we ask the expert all the questions about AI and making music, as it stands today: what can AI actually do? What is it going to be able to do next month? Next year? How do we retain rights? How do you ethically train AI models? Is our work safe? Is it protected? Are musicians done? What about AI bands eating up the streaming money? Drew Thurlow is the former SVP of A&R at Sony Masterworks, Director of Artist Partnerships at Pandora, and part of the A&R/marketing team at Nonesuch Records, and is currently the founder of the Lone Wolves Community - a music business and tech community and think tank, is founder of Opening Ceremony Media where he consults on AI strategy and more to companies such as RocNation, and has a book coming out in 2026 titled, “Machine Music: How AI is Transforming Music's Next Act."
The episode in which we ask the expert all the questions about AI and making music, as it stands today: what can AI actually do? What is it going to be able to do next month? Next year? How do we retain rights? How do you ethically train AI models? Is our work safe? Is it protected? Are musicians done? What about AI bands eating up the streaming money? Drew Thurlow is the former SVP of A&R at Sony Masterworks, Director of Artist Partnerships at Pandora, and part of the A&R/marketing team at Nonesuch Records, and is currently the founder of the Lone Wolves Community - a music business and tech community and think tank, is founder of Opening Ceremony Media where he consults on AI strategy and more to companies such as RocNation, and has a book coming out in 2026 titled, “Machine Music: How AI is Transforming Music's Next Act."
Get more access and support this show by subscribing to our Patreon, right here.
Links:
Click here to watch this conversation on YouTube.
Social Media:
All music written, performed, and produced by Aaron Shafer-Haiss.
[00:00:00] Hey, and welcome to this week's episode of The Other 22 Hours podcast. I'm your host, Aaron Shafer-Haiss.
Michaela: And I'm your other host, Michaela Anne, and we are on episode 118, and this week we are having one of our special episodes featuring our conversation with Drew Thurlow.
Aaron: Yeah. episode one 11, we stepped across the aisle and had somebody from the industry.
join us to talk about creativity and sustaining, career from the other end of the spectrum. If you didn't check that out, it's episode one 11 with j Edward Keys, who is the editorial director at band camp. But more importantly, I.
If you listen to episode 1 0 9, which was our conversation with Major Jackson, the poet and educator, we finally dipped our toes into ai. And after a long time we realized, man, we need to have somebody at the [00:01:00] table for this conversation. So enter Drew Thurlow. he has decades of experience in the industry.
Disclaimer, McKayla and Drew met while working at Nonesuch 2008, they met uh, drew worked in the marketing department as well as doing some a and r before he moved on to be the director of artist partnerships and. Brand relations at Pandora.
Basically, as he likes to say, he took the money that Pandora made and redistributed it back to artists by putting on events and all of that. If you guys ever went to the Pandora showcase at South by Southwest, he did stuff like that. And from there he went on to be the senior VP of a and r at Sony Masterworks.
and now he is. On his own, working independently as a consultant on ai, specifically for, companies such as Rock Nation. he's writing a book on that. He is a consultant for a music tech investment firm that focuses on a lot of technology around ai.
And he just recently started a podcast called Lone Wolf Community Podcast, which is industry and tech centered conversations, which is a [00:02:00] really, a great listen, so make sure to check that out.
Michaela: Yeah. And this was a really in-depth conversation, but also kind of felt like we just barely scratched the surface. So we'll probably have him back on. Keep us updated. As we know, this stuff is rapidly changing, we talked about where we are right now in terms of being music creators and ai, what it can do, what it can't do.
we talked about legislation and legal protections and lawsuits that are happening and also talked about the upside. There's a lot of kind of fear out there about ai and it's going to. Take all of our jobs and make musicians obsolete. And Drew talked about some of the reasons that he would say we can use AI as a tool and that there's an upside to all of this.
Aaron: I was kind of convinced. I'm not like fully sold. I'm definitely still scared and angry, but. a little more at ease. After talking to Drew a little bit,
Michaela: I came away with a feeling of informed optimism because also at the end of the day, it sucks just living in perpetual fear and anxiety.[00:03:00]
So, yeah, let's just keep making music and see how we can hopefully make the best of this.
Aaron: Speaking of making the best of this we invite you into our community so we can make the best of these conversations. And we do that over on our Patreon community. People get events, notice of our guests there are ever changing, ever evolving.
I. Ever coming and going, topics that we talk about. but most importantly, it is the sole way that you can financially support the production of this show, which is quite the undertaking, even with a small show like ours. So if you would like to offset your moral conundrum of using AI by supporting actual artists making actual art, please head to the link in our show notes and support us at our Patreon.
Michaela: And if you are a visual person, this and all of our past conversations are available on YouTube. But without further ado, here is our conversation with Drew Thurlow.
Aaron: we wanted to have you on here with your myriad history in the industry, but had one conversation with Major Jackson, who is like an incredibly renowned poet.
He's the head of the [00:04:00] creative writing department at Vanderbilt. And we found ourselves talking about AI for the first time in three years of this show. And so immediately after I was like, we need to have Drew on here ' cause. can't pretend that AI is not part of what's coming down the pike for us and not
Drew: Yeah. part of like, creativity these days.
it's funny how I kind of got into this a couple years ago is just outta curiosity and now it's all anyone wants to talk to me about.
Aaron: from the outside, like you planted the flag and you're like, Hey, I kind of have an idea of what's going on here. At least you're on top of the crazy biweekly developments. I try to not be like an ostrich with my head in the sand, but I have No idea. I know that like Timberland started a, AI record label that's like. Where I'm at.
Michaela: Yeah. And I'm kind of more of put my head in the sand of like, well, I'm just gonna keep doing what I do. But I realize, honestly, drew, from seeing stuff that you guys have been posting from the Lone Wolves community of like, oh, what are the, legal ways that this is going to impact us, whether we like it or not, as copyright holders, as writers.
[00:05:00] then some of the more. Practical stuff and even like hypothesizing about how it might change us personally and creatively. but if you can kind of share like your background is that led you to being interested in this and kind of this becoming a major part of your work today.
Drew: Yeah, I think it's because I had like an extraordinary experience at Pandora when I was the first music person that the company hired, and
I remember
I had just left Nonesuch and we were super nerdy about our music taste obviously.
Aaron: Mm-hmm. Mm-hmm.
Drew: and for those who don't know me, Kala and I worked together at not such and I took the elevator the first day up.
When I was starting my first day at Pandora and I met a guy who was like, oh, you're new here, right? I haven't seen you in the office. I'm like, yeah, first day. He's like, yeah. I was just at like a Tiesto show last night and I was like, oh my God, I have to work with someone who likes Tiesto. Like, this is
gonna be terrible.
Like, what is this
Tiesto? turns out I had like an extraordinary six years there. And it was, I got to work with music I cared about certainly, but. A lot that [00:06:00] I didn't, but it was because it was like a, forefront of all this new way of streaming.
We pioneered a lot of what we take for granted now DSPs and the way brand partnerships works with artists now. And I felt really good about the advertising dollars. I was steering into artists pockets through content and events that we were doing. And then I
left there and I went to Sony. Sony wanted to kind of create a nun such at Sony, so I was hired to be the senior vice president of a and r at Sony Masterworks.
I spent five and a half years there and I just hated it. It was like everything that modern music is not, and part of it is not anyone's fault. It's just the way that labels and commercial music industry works these days. But when I decided I was leaving Sony, it kind of coincided with this AI explosion.
And I think when Chat GPT launched, I think everyone was like, oh yeah, this is like internet level disruption. And I bought into it, so I started writing about it and thinking about it, and there weren't many people who had worked in the music industry for as long as I had, as an artist and recording for [00:07:00] labels, and then working at labels, and then at A DSP.
Who had the perspective on what this could mean? So I got a book deal pretty quickly. So my book about music and AI is coming out in March. I write for Billboard about it and I have a bunch of clients I consult in the space. I now have a music tech investment fund where we invest in music tech startups, lot of them with AI tools and assists.
So it's just been really interesting and as much as I follow this stuff, even once a week, I'm like, oh shit, this is fucking
crazy. So it's a lot. But I will say my favorite quote is from the woman who runs the Motion Picture Association of America, who said, everyone is so freaked out by ai, but no one's really sure what to be freaked out about. my thesis on this whole thing is actually pretty optimistic for artists in the industry. And we can talk about why, but. This is certainly a pretty big change in technology and how creation happens and
how consumers are gonna listen to music. But it's also like in a lot of ways more the same.
So we can start piece by piece and address like what you're hearing in your community and like what you're curious about. [00:08:00] And there's certain things that we definitely know for sure, and there's certain things that we can kind of game theory out that we don't know.
Yeah. know it's a big question, but let's just start at square one. Like where are we right now with ai? What can it do and what can it not do like in a music? Space, this episode will come out in August, so it'll probably be slightly behind at
yeah,
Aaron: But you know, what's happening right now is like, Velvet Sundown is blowing up
Drew: yeah. Oh my God.
Yeah.
Aaron: seventies AI band on Spotify.
Drew: can we just nod with that shit? the only reason why that's a story is because everyone is like chattering about it. There is a lot of AI generated music out there, so what can it do? It can make music pretty well, it can make
recordings pretty well. And, you know how you know, I consider myself a member of the Americana community.
I was Americana Music Association board member for a long time. And when AI started, really becoming kind of a consumer tool, I. You couldn't really make acoustic guitar. Sounds very well, and I remember reassuring some acoustic. Artists that like, don't worry, like this is gonna be a while. And then like six weeks later, that [00:09:00] wasn't true anymore.
So
yes, AI can make really good quote unquote musical passages or even songs, but
we have a lot of evidence now of more than two years into this that like consumers don't want it and nobody likes it. And the Velvet Sundown, or whatever they're called is not a, real thing. It's like they have all these followers because people are curious about it and they're reading about it.
Of the 250 million songs that are now on streaming services, I estimate 30 million of them are AI generated. And they are like a third of a percent of the royalty pool. it's a rounding error.
No one is listening to this shit. So if you're an artist who's making real music, we can pretty much bet that even though these services can make good sounding songs and musical passages without the context of art or human attached to it, fans have expressed no interest in it.
And that has not changed. It could, but right now that has not changed.
Michaela: Whether that's true or not. What about just the fact that tech companies that are making the AI generated music or Corporate [00:10:00] entity or business, whatever, I can't imagine real musicians making AI music, your motivation has to be to make money, right?
Because it's not like the experience of creating music and sharing it with the world. It's just adding to the, saturation and the noise and the royalty pool. Because listeners also might not understand that like, the way Spotify works is that it's not, fraction of a penny for fraction of a penny like, translated royalties that all the royalties go into a pool and then they divide up among single artist.
And if you get less than a thousand listens, you never get paid out. it's not an even transaction. So like, should that be a concern for working artists who are already facing unjust royalty practices and feeling like how are we ever going to make money fairly from our music?
Drew: Oh, you wanna get paid for your music, McKayla? Oh, I'm sorry. No, that's a fair point. that also has nothing to do with AI as far as I'm concerned. I'm a little older than you guys. I [00:11:00] came up when, in order to make music in the late nineties and early two thousands, you had to like get together with a group of people and spend a lot of money recording music.
Now, anyone can download a free software garage band, and I mean, the, sheer amount of music in the world has been increasing exponentially before ai. And that democratization of creation tools AI's gonna supercharge it. But I think that fundamental issue that you brought up, is not necessarily just because of ai.
and that's when I say that AI's kind of a continuation of what's already happening. Like it's a continuation of all that I am an unabashed proponent and fan of streaming. I. Came up during the file sharing era. I had a contract with Atlantic Records that got shelved. I have a publishing deal with rough Trade.
I never made any money and I didn't deserve it 'cause I didn't put the time into it to be an artist. But there are tens of thousands of people, I forget the latest stat, but there's a, like a, something like 150,000 artists now. Making a minimum of like a hundred thousand dollars a year off their music.
And that figure was like 1,020 [00:12:00] years ago.
I don't think Spotify's fault. I don't think it's AI's fault that there are way more people trying to make music now than ever before. And it's a lot easier to make music before. And as an individual artist, individual person, like you Mikayla and, you Aaron, like, yeah, it sucks.
There's a lot more competition. but I don't know that that's AI's fault. And the overarching benefit is that there's a lot more people like you who are making a living.
Aaron: for me, and like what I hear is yes, there are more and more people making music, you know, as you said, the democratization, easy access to and tools and things like splice and sample libraries and all of that. But it still took like some level of knowledge and understanding of like the tools where
Drew: Yeah.
Aaron: you know, my understanding with AI is you basically just need to be able to write a sentence, and like can get into like, okay, the artistry, especially I had brought up. before, so now you have like, platinum selling, Grammy winning producers that are using AI as a tool and sharpening like the input and all of that. the bar for entry now is even lower. You know, I've
Drew: Yeah, but that,
Aaron: like [00:13:00] loop doctors forever, that just like pile on loops and all of that, but like there's still like some kind of artistry to that
Drew: well, I mean, I think you have to think about Timberland as like a business person and not an
artist. 'cause he's one of the main investors in Suno.
And that is why he started this record label using Suno. But that goes back to my earlier point, everything you're talking about is true philosophically, but in practice what we're seeing is like no one's listening to this
shit.
I had this experience a little while ago where. I was playing something in my iTunes library, I don't know why. And this song came on and the guitar sound was like so sick. It was like mid scooped in the right way. And it had like, just enough gain.
I. And I'm like, it sounds like a tube screamer, maybe a Strat or tube screamer. And I'm like, there's no way that AI can make this like unbelievably great guitar sound. And then I looked at the track and it was from uio and I was like, oh shit. Like
it's, you know, God, this
is like, I was completely fooled is someone who like, considers himself an expert in this, I had it downloaded for another reason, but like no one is listening to this shit.
I know you're worried about the royalty pool, Mikayla, [00:14:00] but like this stuff is not really being factored into the royalty pool. As I said, the latest stats are on Spotify. It's about a third of a percent on Deezer claims. It's about a half a percent. It's a rounding error.
It's not taking away from real artists yet,
Michaela: what about the whole playlist world? in my world and peers, like a lot of what everyone's vying for is playlisting and everybody's lamenting the change of playlisting. How it used to be specific people people knew who it was, who was running the Americana one and you'd like, was very curated and now it.
S you know, more algorithm fed and people are like, I was getting 20 million streams on a song because I got a great playlist, and now I can't get on any playlists. What if all those playlists are just AI generated music? Because people are more and more doing passive listening rather than getting invested in music that they love and picking the songs that they like.
Drew: So far. The DSPs are not incentivized to playlists, AI generated music consumers don't like [00:15:00] it. We can talk about the legal stuff too, 'cause that's important.
But there's laws now in the eu and there'll be, coming here.
That AI generated, wholly generated content has to be watermarked as such, and
Spotify is not incentivized to playlist. AI music and there it's not, it's really it. That's, again, it's such a small percentage of the royalty pool Streaming has its own issues. As much as I'm a fan of streaming. The fact that we've outsourced our entire relationship from artist to fans to streaming is a mistake.
And there's a lot of activity in that area about how to monetize your fans better and how to reward super fans more than just a streaming subscription.
So stru, we need to fix that. But that's not an AI issue. That's a, we relied on streaming too much too long is the only way to make money off recorded music.
Michaela: you know, I don't know the, backend of like the deals. That major labels made with streaming, you and I worked at Nun such And what year did you start at Nun Such?
Drew: 2005.
Michaela: Okay, so I started working full-time 2009, but [00:16:00] I was an intern the year before when I was in college and I left in 2011.
And I feel like even my short stay there, We were constantly hearing about the way things were. Mm-hmm. And
Drew: Yeah. was, almost like being pulled into streaming. And then such is a, specific place because they treasure like, classical music, the avant-garde, a lot of stuff that's not super mainstream.
Michaela: but it felt like we were being pulled into the new era of the music industry. Kind of like kicking and screaming. But also
Drew: Right.
Michaela: of a lot of the, you know, older executives who had been there, had been enjoying, the benefits of the nineties and, budgets were bigger, their incomes were bigger.
I'm assuming. don't actually know that stuff, so I feel like now in 2026 and not being on the label side, but having my own. label deals for my music. and trying to like continually build a livelihood off of my music as an artist, feels like we're now in another [00:17:00] transition and it's gonna be like, who's gonna be pulled into it kicking and screaming and who's gonna be. Smart about how to adapt and just feeling more and more as a musician and an artist, you're constantly the mercy of the changing platforms and you're one, having to spend your time creating your art and trying to evolve as an artist and also pivoting and maneuvering and changing and learning and adapting what is the platform and the way that we share.
Our art.
Drew: that's a good thing to think about. And I also think we have to forget about. The way it was 'cause it's like
so not relevant anymore. I advise these like venture capital guys and these private equity guys in the music industry and they always are like wanting to ask about.
What it was like then, now what it's like now. And I'm like, it was just different. Like were manufacturing and distributing a physical product on a truck to
a retail store and now it's like a media business. and this is where I'm like Proponent of ai, I'm also kind of a fatalist like guys is here. Like this is not going
away. [00:18:00] So
it, it is what it is. But at the same time, like the recorded music industry and like the idea of an artist writing songs and recording them is like such a new phenomenon. And for so long, the companies who were, gatekeepers, who like withheld that and I'm like, I'm happy.
people should have access to creative tools. For 60,000 years, the only way you could hear music was to make it yourself. and a lot of this is informed by like my book research, which delves a lot into this history. But like the whole idea of a recorded music industry where an artist was get paid well that's not even like an early at 20 century thing.
That's like a 1960s to forward thing.
So now all these people are gonna have access to these creative tools. Cool. That doesn't mean that artists and rights holders shouldn't be compensated. Especially if the training data is using their work.
But like
if you're making real art for fans, then thinking about where they live and how they consume it and how you can get value from them is really important.
And it's different for every genre and like what I would advise like a K-pop act and what I would advise you as an Americana [00:19:00] songwriter are like really different. Which
is why these like super fan apps are a tough thesis. 'cause like it's being a super fan by definition is being unique.
So your first a hundred fans are like really valuable to you and important to you. And it you're right, like we're all being forced to reinvent this stuff like constantly. And it's exhausting. And I know you wish you could just focus on your music and your art and your songwriting and that be enough.
But like, it's just not, and that's, just the way it is. And it sucks,
Aaron: we're at a point now where like there's two other branches of this that I want to touch on. One is as a tool, as a creative tool. How can we incorporate it into what we do? a lot of our listeners are also kind of in the Americana community, so like very like.
Band in a room, love music the band, you know? you just said two things that make me kind of want to go another direction, which is rights holders and training data. So
Drew: Yes. Yeah.
Aaron: like the legal
Drew: Yeah. just kind of covered what it can do. So I kind of wanna talk about like what's going on legally, what are the fights to like protect us?
Aaron: I know that like wonderful, wonderful [00:20:00] budget bill that just passed
Drew: beautiful. It's big and beautiful.
Aaron: one of the few bright spots in the thing was there was a stipulation in there that you couldn't pass laws to regulate AI for the next 10 years, which is
Drew: Yeah, it
Aaron: game over. For those that that aren't aware, that was shot down in the Senate like 99 to one. I don't know who the one person is, but what the fuck. Um,
Drew: yeah,
yeah, yeah, yeah. Yeah. Well,
Aaron: that's a plus on that. Like where are we legally, like in the music sense,
Drew: so.
Aaron: as I know, you, you can't copyright fully AI music, right?
It needs human to be able to
Drew: Yeah,
there's a couple things that we know for sure. Number one, I really try to steer people away from the idea that there's an outcome of a situation and that means something. Because like, no, it's a mistake to. Put a, sentence in that bill that allowed state governments not to regulate AI for 10 years, but we really don't know the outcome of this stuff.
So right now there's about 42 lawsuits in the United States, more in
Europe, that are trying to answer the question, can I take anyone's data that's copyrighted and protected and owned? Can I [00:21:00] train an AI model on it? And is that fair use? And only three of the lawsuits are involved around music. Well, most of them are involved journalism and books, but that is the big question of this year, what is training data?
How protected should it be? And then we have all different like really complicated legal threads from there. Like, alright, so say it's fair use. Well what happens if the output makes money? Does. Does that mean that like the rights holders shouldn't be remunerated? If they are, does the tech exist to track that and pay that back?
How do we decide how to weigh a piece of training data in an output? There was an interesting couple of cases two weeks ago on the book side, one was a against an anthropic and one was against meta one was a book publisher, one was a series of authors, and they actually ruled that.
Training data is fair use. Like you can take the history of recorded music and you can train it on a model and it's fair use. That's like at least what one of the judges decide. But the way you get it matters. And what these companies were doing we're scraping the internet and [00:22:00] creating this shadow library of content and that was illegal.
And now there's going to be a judgment, a jury case against anthropic. And if you, while the letter of the law at 150 grand per. infraction, philanthropic could be fine at $1.5 trillion. Now,
obviously like no one's going to, no one's going to find someone like a third of the US', GDP in a lawsuit.
But I think it's also really important for everyone to take a step back and be like training data. It probably is fair use. But like that doesn't necessarily mean what you think it could mean. so many historical examples in my book research my favorite one is when the camera was invented and the first camera studio set up shop in lower Manhattan, and then people could kind of like cheaply get photographs made.
It was a common sentiment that painting was gonna die. Like why would you ever need to paint anything? Most painting back then was portraits.
So why would you need it? And it's true that some portrait painters had their livelihoods disrupted by photography. But what it did photography did was like allowed impressionism to [00:23:00] become invented in flourish.
I could go on and on about how, in a music context, that was true when the phonograph and the gramophone were invented and how people thought solitary music listening was gonna make people crazy, which was like an accepted medical opinion of the American Medical Association in those.
Drew: 1920s and thirties and so taking everything from a bird's eye view, the fact that these tools exist, that we don't really know where the training data is gonna lie, but there's probably gonna have to be some kind of licensing also, like the AI companies wanna license it 'cause they want to get the data in a form, they can use it and there's gonna have to be some way to monitor the output coupled with like, this is a cool tool, it's like a novelty, but fans don't really want it.
They don't want, like an avatar artist, they want a real human. I'm just, I'm bullish on the upside, and I know
all artists talk a lot about like, oh, my music's using training data and they're gonna replace me. Like, there's a lot of upside and let me just talk through a few of them. number one, I think you used Pro Tools,
right? Okay.
Aaron: Pro, Tools.
Drew: So Logic Pro has like generative features built into it. Now you can create a base track, a keyboard track, [00:24:00] and they've had a drum generator for a
while. As someone who like was a session bass player for a little while, like, I don't know if that's scary or awesome, but you can create a full sounding band and get 70 different base tracks in your DAW while you're creating seamlessly not interrupting your workflow.
Setting aside the implications of that for a minute on people who actually play the instrument as a creator, that's pretty exciting. and you're gonna be able to like mix and master like pretty well, pretty cheaply. And putting aside the fact that mixing engineers and mastering engineers, we're not gonna do as many of 'em, like, that's really exciting for a creator.
You're gonna be able to create content for your short films, for your music with some of these really cool visual tools. Like I love these tools. Like I've been able to do so much in my business. I've had so many book proposals coffee table book proposals. I've made content like it's awesome
and there's things coming down the line where you're gonna be able to like to sell subscriptions.
And I fully believe this to fans that want to like remix and reimagine your songs with these tools that are now [00:25:00] in the hands of consumers. And that's gonna be a money making venture for you. And these are all tools that are like, has to be managed, and it has to be served the right way and we have to slow walk it.
But this is the future that I see and it's, I'm pretty optimistic about it.
Aaron: like you said earlier, it is like internet level disruption, and so, one thing that we touch on a lot in this show is like the nearly universal feeling of, whether it's based in fact, you know, and more often than not, it's based in fear that there's not enough pie to go around.
there are limited resources for us as artists that have spent our lives practicing learning, developing skills, developing taste, developing That there's like limited pie to sustain this life you know, to do it at a level, like you need to be able to make enough money to like, spend a lot of time investigating,
Drew: Yeah.
Aaron: failing, learning.
Mm-hmm. That now, like with AI and this like, instantly not slowing your workflow, having 70 base tracks. there's obviously and understandably a lot of fear around that. Like, my reaction is like, oh, that's cool. It's like it's a [00:26:00] novelty.
It's cool. When does that shift from being a novelty to being like, that's just what happens?
Drew: Well, it's a tool I really fully believe that artists and the curation. Aspect of this really still matters. what you just said, like, I, could give you 15 examples of how technology has come along, how people were like, freaked out, thought it was evil. I mean, Not that long ago, in the early 1980s when the poly phonics synth was invented, there was a real sentiment that that wasn't music.
It was a robot.
People shouldn't use it to make music. It's like cheating. It's too easy. I. It's taking hands out of the classical pianist or the people who've really trained and then like Peter Gabriel uses it and everyone's like, oh, well if Peter gabriel's into it, then it must be fine. And now like, show me a pop song that doesn't have a polyphonic s synt in it.
Like,
so many examples of that. So you still need to create something that people want to hear, and all of that you've learned and done and experienced still matters. You're just gonna be doing a lot faster, a lot cheaply. And this is just the more of the continuation. Like I don't think that when you [00:27:00] guys were growing up and the internet started, you saw that as a threat, right? You
saw that as opportunity.
The people that we worked for at Nonesuch saw that as a threat. ' my theory was they're in love with the record business as much as records. they love the business, they love the, tradition of it.
the systems of it and those are things that are going away
now.
and I'm very cognizant that, like, especially you, Michaela, who make Americana music, it's
very different community than
like EDM or
pop music. I think
in some ways it's good. In some ways it's bad. I think it's good because a lot of your fans and the fan base of the music that we both love is.
For people that want a little something, a little bit more acoustic and authentic and genuine. And I wouldn't say that the Americana music fans and artists are necessarily like at the forefront of technology the way that like EDM DJs are who
like went wild or NFTs.
Michaela: And also I mean that's where kind of. My comfort lies is that I think there will always be a population of people who just aren't interested and don't want it. those are my people because [00:28:00] I'm not interested in technology, like I play the acoustic guitar, I've like dabbled with the electric at times and I'm like, but I don't like gear.
'cause I'm a. Songwriter, Songwriter, like I just like songs and singing and the fans that I've gotten to know and what has sustained my career thus far is the real like human to human connection. I don't think that's only for, small communities, because I also see it on the most massive level with an artist like Taylor Swift. her billions of fans are fans of her music, but they're also fans of her and invested in her relationships and her stories and her as the human and even like as a massive mega star that she's known for connecting with fans and even invited fans to like private parties in her house you know, so I think that there will always be a population of people who. it won't be overtaking their lives and maybe we won't even [00:29:00] be aware of the way that AI tools will infiltrate. I think of it almost like as a songwriter. 'cause we, we did talk about, how it could help on the creative front, not on the technology side of creating, but actually on like the musical writing side.
And I've kind of always just been like, why would you use that if you really care about the process of writing? But then I'm like, okay, how often do I go to a rhyming dictionary
Drew: Yeah, totally. Yeah.
Michaela: of a rhyme. Mm-hmm.
Drew: Yeah,
Michaela: some word suggestions.
10, 15 years from now, will it feel like That's how like, I'm like, oh, I can't think of a line. All right. I'm gonna go to ai. Maybe that will feel more acceptable.
Drew: yeah. I mean, I think it will first of all, to go back what you said about Taylor Swift, that is everyone's relationship to music that they love. That's not
just Taylor Swift's
I really want to hammer this point home, like the archetype of the artist is matters and the people's relationship to the context and the stories and the person matters so much.
that, that's why I keep saying that a percent of the royalty pools ai, 'cause like no one wants the shit.
[00:30:00] And, mean, the point of technology is to be a tool and to be seamless and in the background.
And I think as people get more comfortable with this and fans get more comfortable with, this, a perfect example is like Riverside.
you use AI features in this podcast recording. It's seamless. It's behind the background. I don't think it's controversial. I don't think you have a problem, Mikayla using it. It is gonna slowly sleep into these tools that are gonna become natural. And I guarantee that you're gonna sit down in front of a doll one day.
And when I heard that really great guitar sound on uio, it was like inspiring. I was like, I wanna sample that. just think of it as a tool. It's a hypercharged tool, but it's a tool just like any other piece of technology.
Aaron: my fear just keeps going back to. The shadow side of democratization it doesn't take much data to understand that back in the fifties, sixties, when it took, a record label with a budget to get an artist in the studio to record something, baseline, like median level of quality [00:31:00] much higher than what's made.
Today where anybody with garage band can put something out and it's like as an enjoyer of art in general and high quality art, I want the baseline of the art that's being made to go up and you know, to me, like generative ai, the word generative is such a great marketing word because it's not generative, it's regurgitate.
If you ask me. It's not creating new music, it's just regurgitating everything that's already eaten. And yes, you can argue that like us as artists, music isn't made in a vacuum. I'm regurgitating songs that I've heard in the past and all of that, but like. I can make a mistake, I can be recording something and like a mic can be resting on an amp take.
Sam Phillips from Sun Studios, like the first recorded fuzz guitar tone he recorded, and that's because the amp fell off the back of a truck on the radio, his session and the speaker broke and they plugged it in and goes, cool.
Drew: I thought he poked holes in it.
I thought he like poked to
Aaron: But same deal.
You know what I mean? Like, it was a mistake and they went with it. how does AI make [00:32:00] mistakes that stumble upon like happy accidents,
Drew: Why are you assuming, number one, the humans outside of the loop on that process? And number two, I mean, there's this whole thing
about hallucinations.
I mean, I love those videos on like Instagram and TikTok of like really bad AI of like people crashing cars and screaming and like a gorilla interviewing people at a supermarket and like, it's all fucking weird and
broken and hallucinating. And what you just described,
like,
I get it, but it's also like as an artist, you're not out of the loop.
you're using filters in interesting ways. Look at how autotune has been employed over the last 20 years.
this is a tool
the happy accents are very much part of the creative process, and that won't go away.
Aaron: Yeah, and that makes me think of
Michaela: like with recording, we can say like, the technology's gotten so good that you don't have to be. Physically that great of a singer to sound good on a recording, and that can feel kind of like insulting. But then I've read quotes from Linda Ronstadt where she's like, oh, I won't even listen to my old recordings [00:33:00] because you know, I did that in one take and I feel like I could have done it better.
And like what I would give to have, today's technology for back then, how much better Linda Ronstadt could have sounded. like, thank God as a fan wasn't mm-hmm.
Drew: Yeah, I know. Well,
Drew: every single era is a, function of the technology. Like why pop songs three and a half minutes? not because there's some. Innate universal need for humans to hear music in that context. And the other thing that I think is really funny about demonizing AI is like, if you've ever spent any time in like major pop recording session and I don't want to pick on anyone Ka Perry, but like sitting through that shit is like every
fucking thing is sampled in focus group to death.
She comes in, she does a top line they, synthesize that sound signal until it's barely a human voice anymore. And then like it's focus group to death even more. And like then they print it and they send it to the label. Like what's the difference between that and generative ai?
Like, is there a difference?
I mean, there's [00:34:00] clearly a difference, right? But,
like, uh, why are we holding up the old way as some
like golden standard, whereas like in certain contexts where like generative is like not any different to me.
Michaela: So if we have this rapidly evolving, progressing technology that has the potential to be really wonderful a lot of ways, also has the potential to be destructive, I have like equal faith in humanity that a lot of people want, and staying rooted in the natural world and who we are as humans and loving each other and all this stuff.
And I also have equal faith that people are destroyed by greed and power. I just started reading Ezra Klein's book Abundance and it like
Drew: Yeah.
Michaela: with this like imaginative future. If we can use all this stuff for good. so my question to you is, is there anything that the regular people, the musicians, the music fans that we can [00:35:00] do to try.
And help steer without feeling helpless and like the big corporate entities are the way and they're gonna do what they're gonna do.
Drew: Be educated on these tools a little bit like they're not scary
and. you might like them and you might find them really great uses in your workflow. Number two, it's like, to the extent you can influence like your policymakers, I really do think that what the EU has done has been great model.
The EU AI Act, phased in over four years, and it has four levels of ai, interruption from benign to severe. And there's like crazy. Helpful regulatory power to make sure that it's being used for good and that the training data is like diverse. And it's a good bill, it's a good law, to the extent that we need the same one here,
and I really am, this is kind of a non-partisan issue.
the other stuff that's touched on this, like the CHIPS Act and, some of the other. The kind of technology [00:36:00] bills they've really been nonpartisan. Like I think people are taking
this seriously.
Aaron: we're in Nashville. I know a lot of our listeners are around here, like Marshall Blackburn is our senator.
Drew: I know. Yeah. Yeah, I know.
Aaron: awesome on is. AI and protecting
Drew: Yeah. she's introduced the main bill.
yeah, She's someone. multi-billion dollar industry in Tennessee, but she actually like it's maybe the one thing that she's on the right side of. I think
Yeah, it's funny. When I was at Pandora, I did a lot of work with Jason Shavitz, he's like a fox talking head now. He was a congressman in the Bay Area. We did a lot legislation around streaming and copyright laws like 10 years ago, and I just remember being like, man, this fucking guy, and I agree on nothing, but he's on the right side of this issue.
It's
weird.
Aaron: on that, you know, you mentioned one thing to do is like to get involved and use these technologies. I remember this is like early last year probably Justin Trainor the. Massive songwriter that's written or had credit on 80% of the number ones in the last decade.
He got involved really early on with, I'm not sure [00:37:00] which company it was an ai, and sat on a panel of like developing specifically like generative music, ai, and he released a whole statement on it.
He's like, listen, like. We need representation at the table. This is coming
Drew: Yeah.
Aaron: is open. Like it's better that we are involved. And I think that applies to like everybody, not just massive number one songwriters. It's like, get involved share the input
Drew: there's ethically sourced tools out there Suno and UIO get all the attention 'cause they raised all the money and they, they're being sued by the RAA.
But there's plenty of other generative AI tools out there that are ethically sourced helpful, licensing music properly, giving people the ability to opt out.
And then on the marketing and release side, like this is another podcast, but I can talk through all the ways that independent musicians should be using this. Tools for marketing and distribution and royalty collection and box block royalty finding and catalog management.
Michaela: gonna have like a consulting call with you
Drew: Yeah, no, seriously, totally.
No, I will.
Aaron: Yeah,
Drew: I charge Rock Nation to do it, but I'll do it for you for free.
Aaron: [00:38:00] hey, deal benefits of
Michaela: being a cubicle neighbors.
Drew: so I have stayed away from AI and chat GBT and everything because ignorance, fear.
Michaela: And what are the energy and environmental consequences of using AI in chat GPT?
Drew: I knew that would come up. I was betting Aaron would ask that question. because he's the hippie.
Michaela: busted
Drew: Yeah.
Anyway, the whole entire AI ecosystem in the world is roughly equivalent to the global aviation, climate impact. So it's a lot. It's not good.
However, one of the things that AI has over the global flight impact is that the technology is like really good at becoming efficient, and in fact, AI is starting to be deployed in like places like server farms and massive energy systems and finding efficiencies to like a great effect. So if you model out how good this technology is getting and how it's getting, you know, cheaper and better, like kind of Moore's law, there will be a climate impact at some point.
Hopefully it can [00:39:00] be neutral or even. positive, it's trending in the right direction. But I definitely don't wanna gloss over the fact that yes, the AI is about as much of an emitter as like the global flight industry.
Aaron: do you have any way of knowing, like, you know, everybody says oh, every question you ask Chad, GBT is like, dumping out a bottle of water. Is that like an equivalent,
Drew: Yeah. It depends on.
Aaron: every song you make, is it,
Drew: Oh yeah. It's hard to know because
the models aren't really open source and we don't really know depending on the model, they also do things on the backend. Like not all chat GBT models are the same. There's like deep seek, which is kind of a Chinese one built on top of it. Like there's like light models that use less. Someone out there knows it. not me. I,
We'll go do this again in a few months. There'll be a
lot of difference. We'll
probably have some, like settlements. In the lawsuits. We'll have a better idea of like where the legal lines are that will inform the ethical lines. And there'll be a lot more to talk about
Aaron: permanent correspondent? We gotta check in with Drew in the
Drew: Segment,
this segment producer. Yeah.
Michaela: this is, this is Drew's dream besides playing Shaker on one of my records. Yeah, Well,
Aaron: we gotta make that happen soon [00:40:00] yeah, AI replaces them on the shaker. So
Drew: yeah.
Aaron: all Drew, thanks for taking
Drew: All,
spend an honor to be on the other 22 hours podcast
when I'm not burning the world on fire with my ai running servers and queries just to brew the internet when I'm not doing that. The other two hours I am talking to you guys.
So I would love to come back
and, all right. Bye
guys.
Aaron: ya.