Can AI Earn the Trust of Music Creators? (Live from Music Tectonics 2026)
AI is raising big questions for the music industry. Can AI tools support the people who make music? How can companies build technology that respects creators, and what does it take for artists and music professionals to trust it?
This week, we’re sharing the conversation that kicked off last month’s Music Tectonics Conference: a fireside chat between Prem Akkaraju, CEO of Stability AI, and Paul McCabe, SVP of Research and Innovation at Roland. Both are lifelong musicians who’ve gone from working in music stores to building AI tools with creators at the table. Stability AI is building AI music tools for professional creators while Roland, one of the world’s leading musical instrument makers, helped create the Principles for Music Creation with AI.
Together, they discuss why trust has become such a central issue in AI music, the importance of ethically licensed training data, and how AI tools can support professional musicians rather than replace them. They also explore whether AI can open up new creative possibilities, bring more people into music-making, and ultimately grow the value of music.
As part of this year’s conference theme, “Rising Tide: Grow the Music, Grow the Value,” this conversation looks beyond the disruption narrative to ask what responsible AI innovation could make possible for music.
Listen wherever you pod your casts:
Looking for Rock Paper Scanner, the newsletter of music tech news curated by the Rock Paper Scissors PR team? Subscribe here to get it in your inbox every Friday!
Episode Transcript
Machine transcribed
[00:00:00] Dmitri: Let's get started with our first program of the conference. some of you know this, I've interviewed over 500 people on the Music Tectonics podcast, and another several hundred startups in the work I do at Rock Paper Scissors. And I sensed something different that influenced what I was thinking about when we planned the conference and the program this year.
Our theme this year, you might have heard, is the rising tide: grow the music, grow the value. Music tech has been seen as disruptive, and sometimes doing things that aren't the right things for artists and labels and publishers. So we wanted to get everybody on the same page. We did not want people to say Music Tectonics is a place of disruption.
Music Tectonics is a place of innovation, and there are disruptive things that happen, but that's not the inherent idea of innovation. We wanted everybody rowing in the same direction, so that's what Music Tectonics is about, is how do we grow the pie for everybody? I was interviewing an independent publisher named Monica Corton on one of the podcasts, and publishers...
I don't know the publishing industry that well, but certainly they want to make money for songs that are out there, for compositions. But I was kind of interested to find out that somebody that I thought would have a more conservative approach to how they, uh, work in the music industry was really excited and interested in figuring out how AI could work for music.
And I was like, "Well, if publishers are thinking about it," it made me think a little bit like, maybe we should push our thinking around the popular narrative around AI. Yes, there's some crazy stuff happening there, but could we find folks out there who are using AI to grow the market, to grow creativity, to create access to more people without necessarily undercutting intellectual property and copyright and some of those things as well?
So I went searching for AI companies trying to do things right, not ones that say making music is not enjoyable. But those that wanna help grow the pie for music and artists, the ones that were working for professionals, not instead of artists. And I've came across Stability AI that's licensed all their training data, and that's what they're building tools for, for professionals.
And I found out that their CEO is a musician himself and signed on to Roland's Principles for Music Creation with AI, which was one of the first campaigns that went out there from Roland, the musical instrument brand, to say, "Hey, before we get into this, let's figure out what we wanna be doing for humans in this first.
let's think about that first." And I also found that this CEO of Stability AI is friends with Roland's SVP of innovation, Paul McCabe, who's a Music Tectonics regular. So I thought, "Why not kick things off with these two musicians who are pushing the field's thinking on the possibilities of AI in a positive way?"
That sounds good, right? Please welcome to the stage Prem Akkaraju, CEO of Stability AI, and Paul McCabe, SVP of Research and Innovation at Roland Future Design Labs. Yes, that Roland. Please, guys, come on up
Thanks for being here, guys.
[00:02:46] Paul: Thank you.
[00:02:46] Dmitri: We're at the Carousel, at the beach.
[00:02:49] Paul: Cool place.
[00:02:49] Dmitri: Nice.
[00:02:50] Paul: Not Canada.
[00:02:51] Dmitri: Not, not Canada. That's what you all thought when you got here, not Canada. Paul-
[00:02:55] Prem: This is an awesome spot-
[00:02:56] Dmitri: Thank you
[00:02:57] Prem: ... for a conference.
[00:02:57] Dmitri: Yeah.
[00:02:57] Prem: I've done a lot of these, and I'm like, "This is, this is amazing."
[00:03:00] Dmitri: Yes. Sometimes it's the vibe. It's the small energy things. You gotta pay attention to the details and the nuance. Paul, you lead the innovation wing of Roland, one of the most ubiquitous musical instrument manufacturers in the world, and Prem, you build generative AI. Why are you two friends, and why are you on stage together at a music conference?
Paul, you wanna take that?
[00:03:21] Prem: We're besties.
[00:03:22] Dmitri: You're besties?
[00:03:24] Paul: Yeah, we Yeah. Thank you. Okay. So, I mean, it's the typical friend of a friend and, you know, coming in through introduction. Uh, Universal Music Group, I think, actually-
[00:03:34] Prem: Yeah, yeah
[00:03:34] Paul: ... might have been involved in connecting us first, in the AI for Music initiative, which I'm sure we're gonna talk about a few times, today.
Uh, we met for the first time maybe January, NAMM, I think.
[00:03:46] Prem: Yeah, yeah.
[00:03:46] Paul: We did a panel together and whatnot. Yeah. And Prem just happened to drop, "Oh, you know, I, I worked in a music store." And I said, "Wait, what?"
[00:03:56] Prem: I did.
[00:03:56] Paul: I said, "Wait, what?" So, how many people here have worked in a music store before? Bless you people.
Jump and Stairway to Heaven are probably burned into your brains right now. Uh, but in any case-
[00:04:09] Prem: And
eruption ...
[00:04:10] Paul: I've certainly worked in music stores as well. You find out that somebody worked in a music store, it's like, "Okay, how long? Years? Brah." Okay? We're part of the fraternity, for sure. So that, I think that connected us and-
[00:04:21] Prem: remember who was there in the front row
[00:04:23] Paul: Oh, yeah, the owners, right? The-
[00:04:25] Prem: Yeah ...
[00:04:25] Paul: of the store, and they were so proud of you.
[00:04:27] Prem: Yeah.
[00:04:27] Paul: They were taking pictures and videos.
[00:04:29] Prem: They were front row, yeah-
[00:04:30] Paul: Yeah. That was cool ...
[00:04:31] Prem: at, at the talk. And I worked for them when I was 16 years old.
[00:04:35] Paul: I mean, that's the fun stuff. But, I mean, truthfully, and I, I wanna make sure that you get to speak, but, we found out very, very quickly that we are spiritually aligned on the important things around human creativity and making sure that human stays in the loop and stays in control of that loop, as it were, as well.
And, and, uh, and I can certainly say that that's where you are in the company that you lead, is as well, so.
[00:04:56] Prem: Totally.
[00:04:56] Dmitri: Prem, why are you here? Same as him?
[00:04:59] Prem: Exactly. First of all, you got an amazing venue. Again, this is so awesome. This is very special. but I've been a musician my whole life, and that's why I worked at Don Lesman Music Center, actually, in Albuquerque, New Mexico, from the age of...
Actually, it was more like 14, they just let me, all the way to, like, 18, 19, because most people... You know, I wanted discount on equipment. I wanted cost plus 10 because I needed it, at the time, and I just loved being around musicians and around instruments. And so, I think that is a real big difference when you talk about AI in an AI startup because it r- in a s- any type of startup, it really, really, really matters who's the CEO.
Because it doesn't matter so much in a big, established, you know, company. Doesn't matter if the CEO of Coca-Cola really loves Coca-Cola or if he secretly drinks, like, Pepsi. It doesn't matter at all, And because it's more of a different job. It's, like, managing public investors and investor relations and partnerships.
But here, it really matters. And I think it shows because I'm building the product that I would have wanted, you know, and what I would have expanded my own musical palette and talent, because I'll play guitar along with the tracks that we generate, and I'll say, "Yeah, give me in C major," and so I can, like, you know, jam for this, and, like, change keys and actually able to do it.
And so it's really about, making sure that I'm designing that in the essence of producer or artist. And I think that is a really huge difference compared to other people you talked about. 'Cause I actually really like doing it.
[00:06:29] Dmitri: I didn't talk about anybody. I did, I did not talk about anybody.
[00:06:31] Prem: You did not talk about anybody.
You allud- whoever you, you mentioned-
[00:06:34] Paul: I will ...
[00:06:34] Prem: alluded to.
[00:06:36] Dmitri: Well, you know, why is there a trust problem in AI with music?
[00:06:41] Prem: Really because, I mean, we've seen this movie in, before in the music business too. In the music business, bad actors don't really last all that long. They become a flash in the pan, and it's easy to become a bad actor.
You scrape the music, and you kind of go and, try to get a bunch of user, you know, adoption that way. But I've never seen something like that, Morpheus and others and Napster, they just, they end, you know, they don't really, you know, last long term. But the ones who do it right, you know, do.
And doing it right takes a long time. So what we did when I took over the company over, a little over two years ago, is let's do this the right way. It was always licensed from day one. We wanted to continue that path, and therefore, I think we're the only ones with the licenses and the investment from all three majors, you know, Sony, Universal, and Warner, because of that.
[00:07:33] Dmitri: Yeah.
Do you wanna jump in on that?
[00:07:35] Paul: Well, yeah, I, yes. but there's a lot to distrust right now. And, you know, I know we're here to talk about music tech, but in some respects, that's very much a microcosm of the bigger picture. If we zoom out and look at the macrocosm, I mean, you just listen to the news any day, and you're hearing something about an agent escaping a quarantine and taking down a government system or something like this.
so there's the sensationalism around AI and the growing dexterity, agility of, AI. there's a lot of There's a lack of information and a lot of misinformation as well, and, and that's... the lack of information is one thing, and I think probably today we'll talk about some things that, you know, I think we're all working to do to fill those gaps.
The misinformation is unfortunate because it's not just the misinformation of, you know, the telephone tag game, it's the misinformation of companies misrepresenting themselves, revisionist history on their storyline saying, "Oh, we've always been about this," when, you know, those that are following the game know that that's not true.
And so, you know, we've got musicians, music creators that are making choices around tools that they use that you can't even really blame them for making some questionable choices because there's a real lack of information or the misinformation.
[00:09:01] Dmitri: Yeah.
Yeah. Well, we've established now that you're both musicians and you're building for musicians, and that's something you both have said publicly that that's, that comes first when you're building your products, whether it's hardware or software.
How has that influenced how you work with technology and AI in particular? You kind of talked about it, Prem, but where does that philosophy get tested?
[00:09:22] Prem: So think about when text to audio, you know, is many times a very strange process for an artist, right? To sit down and articulate your creative vision in words.
You know? I haven't met really many artists or producers who it's like, "Oh, I'm inspired, let me go write down what I wanna do." No, they just do it, right? And so I think this whole text, we're trying to get away from that kind of input and have inputs where, you know, an artist would either play it or have a MIDI keyboard and we can take MIDI and audio, we support both, or you hum it out where you have an audio-to-audio transfer.
Most artists and producers I know, their first stop is the Voice Note, and we can actually take those voice notes and then orchestrate those and put those in whatever kind of audio-to-audio transfer or style transfer. And so we try to meet the creative process where it actually is, not, you know, articulating it in, you know, creative intent in a text.
[00:10:19] Dmitri: Do you wanna take that one too?
[00:10:21] Paul: Yeah, sure. Um, from our perspective, and so you mentioned briefly the principles for music creation with AI. AIformusic.info, if you haven't heard of these, and there are seven principles that are all grounded on human first, respecting both the process and the outcome of human creativity, and it's very important to include both of those, elements.
But one of the principles, it's the sixth principle, in fact, talks about, the need for companies such as ours to go out and engage with artists and not do what tech companies often do, which is to innovate in a vacuum. and, you know, for... I know for my team, Roland Future Design Lab, we say that no idea is a good idea until an artist tells us it's a good idea.
We have to get it off the... And so we, we build, and this is very counterculture for a company like Roland, where, you know, for, for those of you that own our gear... How many own Roland or Boss gear? Oh my gosh. Thank you. Lily, free hot dogs for all of them, please. Um, uh, no, it's-
[00:11:24] Dmitri: We, we will have hot dogs at our pool party tomorrow, just saying.
[00:11:27] Paul: They're not free, sorry. Um-
[00:11:29] Dmitri: Yeah, they are free.
[00:11:30] Paul: Uh, in any case, you'll know of our brands, Roland and Boss, and now DW Drums and what have you. Hallmarks of these brands are quality, reliability, durability. these are tools that don't break when you're playing them in critical music flow, music inspiration, performance, moments, right?
and we've always prioritized, that, but in order to achieve that, you don't see the works in progress from us. You don't see the functional but not finished stuff because that risks drawing down on these hallmarks of the brands. But that was one of the things that we established as one of the kind of the pillars of Roland Future Design Lab, in order that we can get the voice of artists into the ideation stage at the early stages, is that we built...
We call them... If we're talking to you, we're gonna call it a technology preview. They're prototypes, they're proofs of concept, but we build them and then we take them out to the public. And we take them to NAMM shows and super booths and Music Tectonics, where you're gonna see one of our...
[00:12:31] Dmitri: Show and tell.
[00:12:32] Paul: So that, this is, that's what we call a Project Lydia. and this is Roland's first physical AI product. This is running, a neural network-powered timbre transfer engine that's called Morpho that was done in partnership with a company called Newtown out of Tokyo. that is very intentionally put in a very simple presentation of a,guitar effects, a pedal, a Boss, a pedal.
but then we've taken this all over the world, and we've talked to creators and we've listened. This is actually the second version, uh, which reflects all the input that we've received and we continue to get and so on.
But it's that kind of process that we think is absolutely critical, uh, for developing with AI.
[00:13:12] Dmitri: You know, since you pulled that out There's not a lot of hardware with AI in it yet, right? A lot of it is this software experience.
[00:13:19] Paul: Well, not, yeah,
not, not that you see. Not,
[00:13:21] Dmitri: yeah,
[00:13:21] Paul: not, not- Yeah ... yet. Not, not that you see.
[00:13:23] Dmitri: Let, let me ask you, from that-
[00:13:23] Paul: Yeah
[00:13:24] Dmitri: ... principles of, of AI perspective, why is it interesting, different, or special to have AI in hardware rather than it just working its way in the servers out there?
[00:13:34] Paul: well, I mean, we'll prove it to you a little bit. We're gonna ... we'll set this up for an hour after the fireside chat, just over at Table 19 there, so you can come and try it and listen to it.
It's a totally different experience. musicians are ... You know you are. You're emotional, you're sensitive, you're creative, you're tactile, you're touch and feel. you, you, you work with software, for sure, but your home is, you know, when you're touching your instruments, when you're playing your drums, your guitar.
That's what you build the relationship with. and even modern music creators that are largely PC-based, they'll have a controller that is their thing, that is part of, you know, their creation process. But that's critical, and we've seen it. you know, we've also built some web app, technologies and taken them out and gotten feedback on them.
But, uh, even for those people that are in the know, and, you know, many on Prem's team and, and what have you that have had the chance. We've done hackathons together in different places in the world. and when they touch it, and it's like it just takes on a different meaning and a different place in their creative lives.
[00:14:38] Dmitri: I'm intrigued by it because it feels like when stuff is just kinda out there, you go to a website or you download an app, you're not really sure where it came from and where it's going, and I guess we don't know exactly what's inside the box there. But once you've played with the knobs and things like that, you do get a sense of ownership of it in a way.
You're, physically touching, and then you can record it, you know? Rather than have it spit out an MP3 or a WAV file.
[00:15:01] Paul: Yeah, that, that's a- actually a really good point. And, and I think to validate that, so we've showed ... The first place we ever showed the first version of Project Lydia was a very pocket protector conference called the Audio Developers Conference in Bristol in the UK.
Bristol is the home of, of trip hop, you know, Massive Attack, Portishead, and these were code heads. and they wanted to know what was going on under the hood. That's a Raspberry Pi 5 mini computer, by the way, an 8-gigabyte Raspberry Pi 5. They wanted to know what's going on under the hood. Compare that or contrast that to the NAMM show this past January, where we probably had 2,000 people, you know, try Project Lydia, and I would say that 80%-plus of those people just never even asked.
It's like, "Oh, cool pedal," you know? And we were not trying to be manipulative or, or not transparent. We'll say, "Well, yeah, this is using a new form of AI." "Oh, cool. Can I try it?" and that was the difference.
[00:15:55] Dmitri: So there's not a lot of gear yet, as you said, but Prem, for a music creator in this room who's genuinely trying to figure out which AI tools to trust, what are two or three things they should actually look for or look out for?
[00:16:10] Prem: Well, I know one to trust for sure, is of course us. But I think that, ... But look, look.
[00:16:16] Dmitri: Well, I'm not even meaning like a company. I'm, I'm really like-
[00:16:19] Prem: No, I know ...
[00:16:19] Dmitri: when you're vetting-
[00:16:20] Prem: That's a product ...
[00:16:20] Dmitri: when you're vetting something out there, what should you be looking out for within any product?
[00:16:23] Prem: No, I, t- well, first of all, whether it's ethically trained or not, which we are, of course, I think any user is gonna be looking for a great experience from the product, period. And so as we saw with, like, Spotify and Napster, like if you ... It's not about the money that you're spending, you know? Like, the only thing that beats, like, an unethical or pirated, experience is a, the experience of a licensed platform to beat the experience of a unlicensed platform.
So first and foremost, it's gotta be an amazing experience. It has to solve problems for the user that exists, and I think a lot of times in software development and, AI and product development, people have a hammer looking for a nail. Like, we, I think definitely is defining the problem first and solving it and having an amazing experience.
I think secondly, if you're a producer or an artist and you're looking to commercialize your music, the number one thing you need to look for is that it's commercially safe. And so regardless of anything else, that is, like, the ultimate check is that you can use that output in film, television, recorded music.
You could release it and own it, is the ultimate test, for hygiene when it comes to how that model was trained, what it was trained off of, and the license agreements for example, us. So when you go to ours, obviously, you, you know, you could, y- it's fully licensed, it's commercially safe, and it's royalty-free output.
So you own the output as an artist. I think that part is the most important part.
[00:17:52] Dmitri: Yeah. Paul, there's a version of this conversation where AI and human creativity are in competition. We're trying to talk about something different. You both clearly reject that. What does AI as a genuine creative collaborator, not a replacement, actually look like, say, in the studio?
[00:18:10] Paul: I mean, that is the... $64,000 is not worth that much anymore. That's the $64 million-
[00:18:16] Dmitri: Billion.
It's
billion now.
Just
[00:18:17] Paul: ... billion dollar, uh, question. I mean, it is. and I will say first and foremost that that's the question for artists to answer. What we've heard, and we've actually heard this very, very clearly, is, Tools that, allow the artist to stay in creative flow state.
take some of the administrative work, if you will, out of operating a DAW using plugins, looking for sounds, things like that. So I, I would say that that's one is, you know, assistive, tools that help me to stay in creative flow. and then I think more to tools like Stable Audio and, all the great tools that Stability AI are building, tools that inspire.
And inspire can come in lots of different forms, inspiration for one person can be, "Just give me some, players that I can jam along with." Others, it might be, "Take this idea, and then do this to it." You know? either I'd, "I don't know this style," or, "I don't have this ability, to play in this way.
Take what I've done here and, and take it in this direction for me. Let's see where that goes." So almost more of a, of a muse, if you will. And, and those are two themes that we've heard very, very consistently everywhere in the world that we've been asking these questions.
[00:19:34] Prem: You know, I've never heard a song in my entire life that I loved that I didn't instantly, the next day, and go research who the artist was.
And I think that persists even in the age of AI, for sure. So I think music is one of these mediums that have this parasocial relationship between artists and fans that I don't think is going away. And I think there's a couple things, So I think there's no answer to the question when it's some AI trying to replace human creativity or re- replace the artist.
I don't... So I, I don't think there is a replacement for them because I'm attracted to that music because of my relationship, it might be a parasocial relationship, with the artist. And I wanna know where they're from, I wanna know everything about that. And I think that's gonna persist, no matter what.
And I think there's, a couple things I think about. and number one is that the word of the day, one of the words of the day for the last year has been AI slop, and I'm very encouraged by that. Okay, and that sounds weird perhaps for me to say, but I think it's because then humans can really detect when there's no human at the other end of it, 'cause it's really kind of, it's like nothing.
it's like a wrapped gift that looks really nice, but nothing's inside. So the duration or the durability of, me liking this song or this artist, it goes away almost immediately And that's why we build tools for artists and just enable them to either do it faster, like you said, take some of the needlepoint out of it, but also have creative options they normally wouldn't necessarily have, which is really what, you know, turns me on the most.
And so I think that is going to persist. And when I think about AI slop, and I'll finish here, is that, you know, I say human slop in, AI slop out. Human excellence in, AI excellence out. So I think the artistry plays just as important of as a, as a role as than ever
[00:21:33] Paul: I've told you that I wanna be you when I grow up, right?
Um, you, you said-
[00:21:38] Prem: You make me feel old. That's
[00:21:40] Paul: No, you've said a lot of really important things. That may be one of the most important things I think I've heard you say. Very strongly agree with that, and I think about when we were actually writing the principles for music creation with AI, and one of the, I think it's the first principle that talks about, um, and I'll paraphrase it, but, music with true meaning can only come when there's humans involved in the, in the creative process.
and I'll be honest with you, the day that, in March of 2024 when we published those principles, I was quite concerned that we were gonna get lots of questions about, you know, what is meaning anyway? And, you know, is your definition of meaning the same as my definition of meaning?
And I haven't been asked it once, and I don't think that that's apathy. I don't think it's ambivalence. I think we know, and I think it's your answer. I think, exactly as you say, when we really develop a relationship with music, when it's an album we love, it's a song that we love, we think about the human experience, and whether, whether we're just imagining it, or we, go and we find it.
[00:22:41] Prem: Yeah.
[00:22:41] Paul: And we find out who that artist is, and we, you know, we learn about the story of that artist and where they were born, where they grew up, how they learned to play music-
[00:22:48] Prem: Yes
[00:22:48] Paul: ... who they played with. We learn that journey, and every song that we hear from them kind of appears in different ways in the music that we listen to from them, and for me, that's the meaning.
[00:23:00] Prem: Totally.
[00:23:01] Paul: And AI doesn't have that human journey, uh, to bring.
[00:23:04] Prem: And music has that in particular. Let me just say, you know, everybody here watches television. Name your top three television directors. Silence. No one... We, we have watched 100 movies. I don't care, you know, who the director was. It was a great movie, I move on.
That never happened in music. Eminem, the first time you heard, it's like Detroit from that part of town, 8 Mile, like we all know it.
[00:23:28] Paul: Mm-hmm.
[00:23:28] Prem: You know? And so I think that forms this kind of, this bond. I have a funny way of defining meaning actually too, and actually I say AI cannot produce anything with meaning.
There's nothing meaningful that AI can produce on its own, and again, that sounds weird for me to say, but let's define what meaning is. Meaning is an artist, me making a piece of music and failing 99 times, and then the 100th time getting it right. That's what provides meaning in what I'm making, the song that I'm making.
AI doesn't have that. It doesn't differentiate between 9-1 or 80 or 99 or 100. It just does the thing. But that part of that meaning, we just, we never can actually lose given this new technology. we're making sure that that preserves because I think that's the thing that's gonna make something meaningful.
[00:24:15] Paul: And, and that's why the principles say we need to respect both, the outcome of music creation but also the process of music creation.
[00:24:22] Prem: Yeah.
[00:24:22] Paul: And, and I mean, if you've, and I'm sure most of you have tried the prompt-based generative AI, I call it the elf yourself music making. Right? but it's, you do it and you, you marvel at, you know, what you're able to do and it's like, "Wow, this is incredible that, you know, I was able to kind of conjure this."
But if you've been through the process of learning to play an instrument or learning to write a lyric or coming up with a chord progression and you know what that journey is, and yes it's difficult, but that is the experience of music making. and again back to meaning, when we hear music that resonates with us, we just know-
[00:25:01] Prem: Yeah
[00:25:01] Paul: that that process, that pro- sorry, process process, is in behind everything.
[00:25:05] Prem: We understand
your Canadian accent.
Thank, thank you. Thank you. Eh.
[00:25:11] Dmitri: So I'm gonna speed ahead a little 'cause we're, we could talk for hours probably but we also want people to try out, I mean, Stability AI is right inside. You can demo.
[00:25:19] Prem: Definitely have to check it. We're running everything locally too, so it's not even... Talk about like conforming to an instrument-
[00:25:24] Dmitri: Nice
[00:25:25] Prem: ... or hardware. Everything is local.
[00:25:26] Dmitri: That's good 'cause, you know, Wi-Fi. But-
[00:25:29] Prem: Exactly
[00:25:29] Dmitri: ... and then, and then, and Paul will be inside with, with Project Lydia but let me ask you a couple more things.
Here's my rising tide question. I started the conference with the belief that the right technology grows music. More artists, more revenue, more creativity, not just a reshuffling around of what already exists. Does responsible AI actually grow the pie? Or does it just protect the current one?
[00:25:53] Prem: Well, every technology...
by the way, technology always, it develops in advance of copyright, right? That's why there's always this, you know, gray period in everything from streaming music to now, et cetera. but always after the noise it always increases, whether it be increasing distribution or new instruments, et cetera.
And I always think about... And there's always a fear factor too. There's always a fear factor that they say, "Oh, well, this is gonna displace..." Like let's talk about the LinnDrum machine. That's gonna replace drummers, right? And then they always said the same thing after that which is, "And this innovation's different than the other one."
Just like when we talk about AI. Every single fucking time we say the same shit, okay? But think about the LinnDrum machine. The LinnDrum machine only made 15 sounds and it changed music forever. We wouldn't even have a song like When the Doves Cry by Prince without the Linn drum machine, 'cause human drummers don't play that way.
So this is an example of how it grows or how it created something. It wasn't like he needed to do the drum line faster, it was he wanted to create something that couldn't be made before and I think that proved it. So I, I take a lot of inspiration from things like that, when I look at AI as well.
It's, you know, we always say it was bigger than the last, but at the end of the day, net-net it always increases.
[00:27:12] Dmitri: Wh- when I look at that whole rising tide theme and I think about the direction of what AI could be if we get it right, I start saying, well, is there more revenue coming into the industry?
More revenue coming for artists and labels because of the licenses and because of what gets produced? Does that grow the market somehow with it, with a new genre?
[00:27:31] Prem: It's a b- I, I pray that there's a new genre that's gonna be made and a new instrument that's gonna be made, from AI. because those have lasting...
You know, that's the stuff in the music industry that really lasts. But I think there's no better time to be an artist right now because for responsibly trained AI companies because you get, There's two ways to win. You get paid on the input and the output. So basically your input is being licensed by companies like ours, and you're paid a license fee or participation, and on the output you're being able to make, like, so much more, you know, of your music and so much more creative expression that then you go and monetize which is fantastic.
[00:28:07] Paul: And, you know, back to the core of the question. does, is it going to, you know, grow the TAM? Yeah. Uh, grow the total addressable market. As I'm sure we probably all know, music and technology have lived in this synergistic relationship since the very first note that's ever been sounded.
And there have been many, many locally disruptive technologies that have come into the world of music that ultimately, on the other side of it, as the dust has settled, has grown, the number of music creators, the number of people that enjoy music. and very much, I'm Canadian, I'm naturally an optimist in, in this.
I'm, I'm a Canucks fan, so I'm a real optimist. but, very much believe that done responsibly, and there are motions that suggest that more are making the right choices. the road's gonna be bumpy to get there, but it is happening and things are slowly moving in the right direction, yes.
The one caution, and this is why AI For Music came into existence, it was like this existential crisis pang that I felt, that day in November of 2022 when ChatGPT dropped and it was like, "It can do what? hang on a second. Let me just start to get my head, around this," and I couldn't. But the one difference about this disruptive technology...
So, people that know, sociologists and whatnot, will talk about AI as being a general purpose disruptive technology. So we have locally disruptive tech that influence a sector, an industry, a segment of society. Then we have the general purpose disruptive tech that influences and touches everything, and that's where AI is categorized.
So one, we must not underestimate the scope, of this disruption. and even if you haven't necessarily felt it permeate all the corners of your life, wait a month, right? It's, it's-
[00:29:52] Prem: Mm-hmm
[00:29:52] Paul: ... you know, every, every moment that passes. But here's the other one, and that's where human agency sits in this equation.
So when we think about the technologies that have disrupted music-making, and whether that's sampling and digital recording, sequencing, MIDI, electricity and the electric guitar, Les Paul, thank you very much, multi-track recording, and on down the list, right? In the process and the outcome, human agency has been present.
Yeah. The risk here is that done irresponsibly, human agency is removed from the process, and that's the existential threat, and that's why there are initiatives like AI For Music, aiformusic.info. Please, if you're a patron of brands that you don't see listed on the almost 200 brands that are on that site, ask them what they believe in.
It's okay if they find another way to communicate what their principles are, but make sure they're communicating their principles. And we'd appreciate it if you asked them to join AI For Music as well because it is a rising tide, does lift all boats there as well. But that's the concern, and that's what makes this moment different and why there is risk that can't be ignored.
[00:30:54] Dmitri: All right, we've got time for maybe one question. It's like one, one minute left, and I've got one more question I wanna ask you. We spend a lot of time debating whether AI is good or bad for music, but what if that's the wrong frame? What if AI is just a new interface like the electric guitar or the synthesizer?
What does music look like when we stop arguing about it and just start playing? So just, we got about 30, 40 seconds each, but where do you wanna go with that?
[00:31:20] Paul: well, first of all, more people, getting to the point where they can play together sooner. because I think anybody who's played in any kind of an ensemble or just with another musician, you know that there's the solo journey of learning to play an instrument and trying to build up the confidence to let somebody hear you play, and then there's that switch that clicks into place when you play with somebody else and you experience that joy, for the first time.
So there's a, there's a more people that are enjoying playing and collaborating, together, I think is, definitely going to be one thing. I think we're going to see, a continuing resurgence in the importance of live music, in life as well. And we have to remember, right? like the shift in the fulcrum in our industry to recorded music as the kind of the center point of consumption and commerce is less than 200 years old, right?
the turntable's about 155 years old. That's it, right? and so that's not very long. The scope of human history, that's not a very long time, right? So, I think we're going to see a continued resurgence in how we enjoy and create live music experiences.
[00:32:26] Prem: I mean, I talked to a, really famous producer and they were talking about AI and how much they loved it because they felt like artists and producers need a kick in the ass.
Like, we need the next kinda thing so we can then you know, take our creativity to another, another level. I love that, and I also... Interestingly, one out of 10 people in America play a musical instrument. , It's a huge, you know, population actually, when you think about it. wouldn't that be wonderful if that can actually be instead of 10% of the population, 15% or 20%?
Because now you don't necessarily need to learn music theory as much as you did, or have the dexterity, you know, play actually an instrument, you know, but it can accompany that. , You can assist those one out of 10 and also grow that market as well, because I think it's a really important type of medium, as we discussed, for self-expression , in a way that other mediums don't have that.
it's very vulnerable and it's very personal.
[00:33:21] Paul: and my kids will laugh if I say the dollars to donuts. He got it. Um- I say that to my kids ... the, so the 90% that don't play instruments, if you walked up to every one of them and say, "Would you like to play an instrument?" They would-
[00:33:32] Prem: Yes
[00:33:32] Paul: ... all say yes-
[00:33:33] Prem: 100%
[00:33:33] Paul: 100%. And that's the opportunity.
[00:33:35] Prem: That's right.
[00:33:36] Paul: Yeah.
[00:33:36] Dmitri: Amazing. Paul McCabe, Roland, thank you so much for being here. Prem- All right ... Akkaraju, thank you for being here. Yeah. All right. Thanks, guys. Don't forget your Project Lydia over here.
Let us know what you think! Find us on LinkedIn, and Instagram, or connect with podcast host Dmitri Vietze on LinkedIn.
The Music Tectonics podcast goes beneath the surface of the music industry to explore how technology is changing the way business gets done. Weekly episodes include interviews with music tech movers & shakers, deep dives into seismic shifts, and more.
