How Artists Can Shape AI Before It Shapes Them w/ Imogen Heap
62m 5s
The transcription discusses the rising trend of AI in music creation, with artists like Zania Monet using AI tools to produce songs that are climbing music charts. Imogen Heap, a Grammy-winning artist and technologist, shares insights on the impact of generative AI on creativity and the music industry. She mentions her AI voice model "AI-Mogen" and her work on oracles to protect artists' intellectual property in the age of AI. The conversation delves into the need for identifying and distinguishing between AI-generated and human-made works to ensure fair treatment and recognition for artists. Imogen Heap also touches upon the evolving music landscape, where major labels may start signing AI talent and the potential for AI to push creative boundaries in music production. The dialogue reflects on the exciting possibilities and challenges posed by the integration of AI in music creation.
Transcription
11636 Words, 59137 Characters
When AI becomes the next pop star, who gets the spotlight, the artist, or the algorithm? You might have caught recent headlines about so-called AI artists or artists that heavily rely on AI tools like image, voice, or music models to make their songs. There's one artist, Zania Monet, for example, who recently signed a record deal and is climbing the gospel charts and is playing on radio stations across the country. Meanwhile, last week, a bunch of songs in the top 10 on the country sales chart were also songs that stemmed from this kind of AI artist. At a time when AI generated songs seemed to be exploding, cropping up everywhere, from TikTok trends, the Billboard charts, it felt like the perfect time to dive into how this will impact artists and what can be done to protect human creativity in the process. And I couldn't dream of a better guest to join me here today to discuss this topic than image in heap. Welcome back to On the Record, a music business podcast from Billboard and Sickbird Productions. As always, I'm your host, Kristen Robinson, and I'm thrilled to have image in heap on the show today to talk about the rise of generative AI and how it will impact artists. Image in is a multi-grammy-winning singer, songwriter, producer, engineer, and technologist who has been making music for over two decades now, and it's clear that she still is somehow a few steps ahead of the rest of us. We will unpack this growing AI phenomenon and what these tools mean for the future of creativity. As Imogen herself recently put it, the Genie can't go back in the bottle. And after years of research and development, she's finding ways to use AI ethically within her own artistic practice. There are really few artists to understand the intersection of music and technology quite like Imogen. Her latest singles, I Am and Afterlife, for example, both use her regular singing voice and this AI voice model, which is trained on hundreds of hours of her own singing, which she calls AI-Mogen. She's also been hard at work developing oracles, a system designed to help artists manage their IP in an age of AI where deepfakes and training models on music without permission is very commonplace. We'll also touch on the 20th anniversary of Imogen's influential album, Speak For Yourself, which she released at another time of great change in the music bives. Back then, CD sales were down, MP3s were taking over, and almost no one was having an easy go of it. But still, Imogen managed to release the project independently and to great success. Today, the album is considered a modern pop classic, and ever since she's gone on to serve as an inspiration for music's biggest stars, including Taylor Swift, Ariana Grande, Olivia Rodrigo, Pink Pantherus, and yes, even Jason Derulo. In the years since, Speak For Yourself's release. All this to say, I am very, very excited for our conversation today. So let's go ahead and bring her in. Imogen Heap, thank you so much for coming to On The Record. I'm so happy that you're here virtually from London. Thank you, very happy to be here. Where did you just fly on from? You were just on a plane like two hours ago. I was coming from Lisbon, lovely city. It was actually torrential rain. I didn't know if I was going to get back here because the thunder and lightning last night was so incredible. I thought that there was somebody like going over my hotel room with a massive car or a big train or something, but it just turned out to be thunder. So yeah, we might have, but we've stuck on runway forever. It was your own night. There's nothing worse than trying to get back quickly from the airport. So I really appreciate you still being able to make it. I'm really excited to chat because I think that you are one of the artists that really just has been keeping your finger on the pulse of what's going on in the world of like cutting edge technology, generative AI, all that stuff. And so that's one of the things that I report on at Billboard. It's become a huge focus for me over time. So before we get into that kind of a chat, I wanted to kind of throw it back to the beginning of your career when you're getting started. What were some of the technological tools in your studio back in the day that really inspired you? Well, I didn't have a studio in the beginning because there was very expensive to have a studio. But I did have a keyboard, I had an eights and in Sonic TS 12, which I still love. And I did have access to computers. So when I was 12, I started learning on an old Atari and there had some software on there at school that nobody knew how to use. And it was in like this cupboard. It was basically like the naughty cupboard. But I realized later in life that my head, the music teacher was just kind of being nice to me, just getting that cut. But he knew that I just wanted to make music on the computer. So that's when I started realizing that, oh wow, there's something in this computer thing that's like helping me be more than I can with just writing it down on the paper and hoping that an orchestra might play it one day, which they never would. Well, they did eventually. So yeah, it was like computers. And then when I was 15, I went to the Brit school. I was a second year in this school in the UK, which lots of musicians go to because it's the only one of the only schools like it. And that's when I learned to use Cubase. So I did know my way around to do that, but I didn't have that much equipment. I had a QI-70, but that was a bit later. That was like when I was 21. So I could make little beats on that. Do you think that like technology has always inspired you to start writing? Or what do you think usually kick starts the songwriting process for you? It really is more about time. Like what am I doing? Any musical instrument anywhere or if there's any time where it's like, oh, actually, I could be doing music now, then it really doesn't matter where I am. I can make music out of anything. And I think it is really about just feeling calm and excited and like there's not a lot on your back right now. And the sense that you actually might have time to finish something or even just manage to do a verse or even if you did capture something, I'll have to put this fun little. It's called the drone machine and it's made by style phone. And I love it. And I spent four rounds just like nudling around on it on the plane the other day, but I didn't have a way to record it into my iPhone. So, and I didn't read my laptop. So yeah, there's just, it's right now it involves lots of steps to just kind of clear the area to make me enable me to just actually put something down. But most of the time it's a note pad. I just write all the time in my little note pads or in my notes. When I was younger, it was a piano. It was a piano or it was a cassette player. And I would just come up with something. But now life is different. Very hectic. There's so much stuff going on, there's little piece. So it's just little scraps of things really here and there. And like this little style of own things, really cool. Interesting. So I want to get to AI-Mogin oracles and the Nimeo gloves. But before I get there, I am very curious like when was the first time that you realized that generative AI was going to be very impactful in music? I don't know, I think it like weirdly crept up at me. The first time I did any generative anything was with a human called Rob Thomas. And this was about 15, 17 years ago. And he wrote code for apps. And there was this brilliant app. Basically it would take the world around you and it would synthesise it in harmony. So if you're on the train, it would kind of have resonant frequencies of everything and then it would just sound all in time and in tune. RTA DJ, that's what it's called. And there was all these different scenes and they were really great. Anyway, I loved it so much. I reached out so I was like, can I speak to somebody there? And then he came out and we did this experiment where he would he would personally analyze the way that I improvised. And then he created a little program so they could generate music in the style of the way that I played piano. So he was just like, you know, mentally creating a little algorithm as a human, as kind of a portioning kind of appropriating what he thought was me. And it just kind of occurred to me, wow, that'd be so useful to be able to just like start improvising. And then the piano continues on without me. And then I can kind of jam kind of with myself, but like a one step removed, but it still kind of me. And so I always was really inspired by that. And that started a song called Run Time because we were going to make a running app out of it that would like work to the rhythm of your running and everything. And I think that that's a total thing now. But we wanted to we wanted to explore with that. So it was yeah, about 17 years ago. But I suppose the way that I feel now it's impacting is just everybody's terrified. That's not true, not everyone's terrified. I'm not terrified. I'm just aware that even not just even music, everything, everything is is being turned upside down. And it's not all a loss. There are many things that we can do to mitigate it and to set a standard to to make things better for us in the future by getting things right now. And I'm excited about the aggression. And do you see it as an evolution of us as people? And I'm excited to with all the possibilities that could happen, hopefully go and talk about that for about four hours. But I felt like only really everywhere else I felt the big change in the last two years, really. But I've been doing like AI judging of AI music competitions for about, I don't know, I've been asked to do that like three or four times over the last 10 years. And there was then a sudden like jerk basically of like, oh wow, this sounds completely amazing compared to what it did then. And you know, that is incredible, you know, the shift that really happened over, you know, yeah, basically. I feel like for me, when I noticed a huge shift was I started reporting on AI music and it was just like they could generate kind of a shitty but passable instrumental. But then when Suno came out and you could really generate everything at the click of a button and it sounded pretty convincing and increasingly it's gotten even more and more convincing. That's when I really felt felt like, oh wow, it's suddenly here. It felt like it just a moment when things really did shift. And it's interesting that we're talking this week because it feels like right now a lot of these so-called like AI artists, which I don't know how I feel about that term. But there's a lot of people who are using AI tools very heavily in their creative process that are now actually seeing some level of success from it. Because I think at first, it was maybe a creative experimentation, but they weren't rising up the charts or they weren't trending on TikTok or getting a lot of streams. But now we're starting to see that with like a Zenia Monet. She has made a lot of headlines recently. I'm wondering for you as these kind of heavily AI-generated projects start entering the marketplace alongside human-made works. Do you think that these things should be treated similarly in terms of what royalties they make, what opportunities they have? Or do you think there needs to be some distinction between AI works in human-made works? And all the gray in between. Yeah, and then all the gray in between. Yeah, for sure. No, I do think we need to have a way to identify in some form, however little or granulated that can be right now. But we definitely need to build the layer that allows the humans to show their input and to be verified by other humans as that is that person's work. And because I trust this person, to trust this person, and that's what we're doing, which we might talk about later with oracles. It's just a place for a human to say this is me, these are my works, these are the people that I work with, these are all the contextual information around this piece of music and how I made it and the instruments and whatever it is, whatever we want to find out. I'm creating this wealth of information that the future and now industry can build on. So until that happens, we are still kind of floating around in midair because we don't even have the bedrock of understanding even right now as we're still in this kind of traditional reason-making way of identifying human work even now. And that's one of the main problems. Yeah, that's something that I've been encountering in my own reporting is people will come to me and they'll say, "Hey, this song is AI-generated." And I'm like, I don't know for sure that it is AI-generated or if it is fully human-made. And I think AI detection tools still have a long way to go. And I'm really, really hopeful that there's some strides made in that because right now it's one of those, it's almost like a guessing game and you'd hate to accuse someone of using AI or say an AI work is human and have that be completely wrong. But like you were saying before, there's like such a great area in the middle. When we talk about or we see headlines about AI songs or AI artists, this can be anything from something that is fully generated just using a prompt and then the human steps back or it could be something where it's mostly human-made but there's just an AI vocal that's incorporated. Christian, actually, I just want to butt in because you said then about mistaking human music for AI music and actually embarrassingly, I did that the other day and I've got really bad about it. So I discovered this artist when I was listening to a playlist and I discovered this playlist because of my new favorite artist called Blowen, which I know I'm a bit late in the day, but I just loved this record that he's done. So it recommended, you know, go listen to this, other stuff. So I found this long playlist and I listened to loads of stuff and I thought, oh quite lame notes on this plain journey. But one of them was that this artist called Data Cluster and I was like, wow, this really sounds like Max Cooper and John Hopkins, like it really, really does. I was really suspicious and I looked up on us, you know, on the biography about this person and there was it just said data made, oh, is it made with errors or something like that? And I was like, this is very suspicious. And then I looked this person up and I did actually find a person behind a but I thought, ah, but that could be a reason. Anyway, I found that Instagram page and then I followed them and this person was like, oh my god, imagine, hey, is this real? Are you real? I love your music. And I was like, are you real? I'm just going to say like, if you were, because I just said this record that you've done, I absolutely love it. Like I'm really, really enjoying it, but I can't help feeling how similar is sometimes to Max Cooper and my friend John. And he was like, oh, thanks so much, he was like, by biggest heroes, that's like the best thing he could have ever said to me. And I was like, no, and I don't take this the wrong way. But did you create a model that kind of, you know, matches them up together and then generates them and kind of just stop over the top, who's like, absolutely no way. I can show you the stems, I can do this. And fair enough, we got like, you know, a bit offended. But it has got to that point where we can recreate things and honor people and we do all the time. Sometimes without knowing it, we just, we go, oh, come up with this idea and actually it was something that some of the else did, but you just forgot about it and you thought it was a original idea. And when it happens all the time. So I'm quite, yeah, I, we definitely are that place where it's very confusing right now. And we do need some clarity, but we are, we can do it. There are, there are tools and everybody wants it. So I feel like the silver lining really of this dark cloud as some might see it as AI music taking over is that we are going to get the data layer of works. We're going to get this complete data layer of works because people will want to prove not only that they're human, but they want to go that they want to actually just say, no, I'm human. And these are all of my works that I actually contributed to, even if I'm a mixing engineer or whatever it is. So I feel like, you know, this panic in, you know, how are we going to, how are we going to know what's what is also going to empower us to just do what we've needed to do for like a hundred years? I recently spoke to someone at Hallwood Media, which is signing AI artists. They've signed this Zenia Monet artist, which has gone pretty viral in recent weeks. And one of the things that they said is because Zenia is backed by a real person who essentially writes poetry and then uses, you know, to help complete that poetry into a song. They record, or they have her record herself making her music so that if anyone ever has questions about it, they can actually show like, here's exactly what she did. Here's exactly what AI did. I could kind of see that being something that becomes a thing in the coming years of just trying to create a log of, here's what I did. Here's what AI did. Is that increasingly becomes part of the creative process? But I do want to go back to that point about Hallwood signing AI artists. They're the first label that has been pretty public about doing this. I'm wondering in the coming few years, do you anticipate that this will become a thing where major record labels are signing AI talent, or talent that heavily uses AI? And how do you feel about that? Do you think that that's right? I think lots of things when he said that. I feel like a lot of major labels are signing music that sounds AI generated to me anyway. It's just like, oh, that just sounds like that last thing and that last thing. Nothing's changed. So it doesn't surprise me at all that that would happen. And wherever there's money, obviously they'll go, if they think they can make money out of this artist. I think right now it's exciting. It's such a nascent space, exciting kind of beginnings, where it still sounds a bit AI, and it's like, you're not quite sure. That will soon disappear. We absolutely can't tell. I actually love these early beginnings where the voices just sound like how they're jumping around or the performance of the voices isn't quite right, and the diction of where they say things and their accents is all funny. I really like that. I'm kind of going to miss that. When AI and humans create music that we never could have expected, you know, that just like really we just wouldn't have gone there because you couldn't really imagine mashing that thing with that thing because it would have taken so long to do it with a band, with a fit. I don't know. There will just be able to iterate and hear things differently and it will really just like make us hyper creative. So I'm excited for the music that's going to come and the people that are going to mash and hack these systems together to create like weird kind of weird sculptures with small heads and a butt music, because why not? We're not really pushing the boat out a lot, and I really think we need to. I think we'll go through this boring phase of people just mimicking what we do, and then once we've got over that, we will create music that's like really astounding and exciting, but it will be a complete cross between us. What you just said earlier really reminded me of a quote from Brian Eno who I'm going to butcher it. I don't know it off the top of my head, but he talked about how after a technology gets really good, people start to have a fondness for like whatever the quality of it that everyone criticized when it first came out. I mean, early photography and film did not look good. There was a grain on it. It didn't look totally real to life, but after a while now we're adding film grain onto things to make it kind of emulate that style. I think it'll be interesting to see what people think of looking back on, I guess, like 2021 AI-generated images, where the hands are all messed up, and they have different fingers that are not right. But yeah, it'll be fascinating, but I wanted to go back to something that you've referenced before. You said that even 15 years ago you were working with AI to try to create an algorithm that mimicked the way that your creative process worked. I was using a human to do it. Using a human to do it. Okay, yeah. But it was the beginnings of this generative idea that you can extend yourself beyond just your human capacity that you can create a code that will emulate you and beyond. Yes, and so now, like you've recently released a 13-minute song called I Am, which also has a beautiful visual that goes along with it. And in that song, it's not just you singing, it's also AI-mojin, which is your essentially digital self or your AI-self. So can you tell me about what AI-mojin is and why you think that that is an interesting creative interplay to essentially collaborate with yourself? Well, I don't think the way that I've used AI-Imagine in this song is particularly creative. It's kind of a very long-winded kind of silly way around to do things, but I did it as a statement, and I did it to rile people up, I suppose, and just have, I'd be able to have this conversation. Because the song in the beginning, it takes you through this journey, really, of what I've experienced over the last four years. I recognize, so I was starting to think about, who am I? What am I? What, you know, this ego? What is it? What's stressing me out? Like, what is it when you get rid of all of that in just in breath or in silence or in the noise? What's there? And then I started to think about AI, because what is that that the AI isn't? What can I feel that is something that an AI can't feel? And I kind of got down to these bits, these tiny bits. And so the noise section in that song is like the annihilation of my ego. I mean, I'm not saying I haven't gone ego. So then I needed, I needed after care after that. I love this idea that AI is our child. AI is something that we are raising together as the mother and father, as the, you know, as the adults raising this child. And what I'm actually doing is of that. And right at the end, there's the AI voice of Mojin. I wanted it to be an AI voice, even though I had to sing everything. And so I sang all the parts and then I put it through my AI model so that it's the model of my voice that's then singing the words rather than my actual voice and I sang. But the melody of it and all the, you know, inflections and all of that is me. But the sound that it's like changing the sound on a keyboard. Yeah, so it's like putting a filter on top of it in a way. So you still had to do the rest of singing. Yeah, I had to do singing, but it's like changing out the sound for another sound. So it could have been the trumpet. And I wanted to trick people, you know, into going after all of this music, after all of this, like quite traumatic noise section, that they would, they would feel something and that voice would not be my voice, it would be AI, Mojin's voice. I wanted to create a discussion. I wanted to, you know, show people that we already don't know. But does it matter, you know? I mean, I say in the title, it's AI, Mojin. But it was all ethically sourced. It's all done in there. The best way possible. And it's my own voice. And I didn't use any, I didn't generate any music. Of course, everybody has already, some people have already cancelled me for, you know, even talking or saying that AI is in my music. So many of them, I can't believe the image and I'm never listening to music again. You used AI to generate the soft. I was like, no, I did not. I wish that I could have done it, because it took me four years to do it with a hundred thousands of hours. Anyway, so my point is that everybody is fearful of it. But we can still feel, you know? And what, what is art, you know, art is something to someone but not to somebody else. But if essentially, it speaks to you and it makes you feel something, does it matter? If it makes you recognise something, you know, self, I mean, essentially, or AI is generated from human work. So, and I think that what we'll end up with over time is the AI models that we do rely and live with will teach us that we don't need anything else. We don't need this capitalist horrible world that we have got to the extreme of. We just need each other to be kind to each other for communities to care about. You know, we don't need happiness out there. It's not something we can buy a Porsche, something we make us happy. You know, so the voice right at the end of the song is actually my daughter's voice. And she's the, she has the last word. And that is to say that, you know, AMOGen, Scout, AI, there are children. And we need to take care of them and we need to do a good job of showing them how to do things, the way that we would like them to be done, not how we currently doing a lot of things. That's so fascinating. I mean, I love that you've been able to make this model with AI AMOGen that is totally within your control. So, you can do with it what you like. I think the thing that gets scary is when, you know, anyone can manipulate or find anyone's voice model online or train a voice model based on interviews that people have done in the past and just run with it. I'm wondering in an age of AI, I wonder how artists will continue to self-curate. I think such an important part of being an artist is having the taste and the curation to know what you want to say. What songs you want to put out, what songs you want to keep for yourself, what interviews you want to do, what things you don't want to do. In an age of AI, it feels like you're losing a lot of control over yourself. Do you have fears around that? I mean, we already have lost control. People be quoted as constantly, you know, say that we've written something when we haven't, don't credit us when we have been a part of it. That's still very much there. So, I do actually think this is going to really, it's just going to force us into creating something that will make sense of what we have already and for the future so that we can put a flag in the sand as humans and go, okay, up to this point, it was human-generated and then there was like this blurry space and then we figured it out and now you can really discover what is what. I'm excited because I like, I have you, other people have also used my voice. There's an artist called Beretta who's created a great dance track with it and there is an artist called Lucy Hayes, it's done a really amazing song with it too. And you can go to, I have a profile on DSPs as AI.Mogin and you can see the songs that she heard voices song, including my voice, including my song. But the nice thing about that is, as I say, you have control. Again, I think it just comes down to this like core missing piece that we don't have, which is an ID layer like identifying home for each individual that I think we'll just have in time anyway just as much as there's one physical version of you in the world, even though we'd like there to be more of us. And there's one of me, but there's no digital one version of me somewhere that I own, that I can, that anyone in the world can go to, to find the facto information about me, you know, up to the minute if they need to, so that I can express whatever I want to express in like a kind of umbilical core to anything that I might be connected to in one instant, from one place. I want to be able to do that. I'm wondering, how are you seeing some of your peers use AI in their musical process that inspire you? Is there anything interesting you've seen recently? I'm not seeing any of my peers doing any of these with AI. Maybe I've got the, I don't know, I just feel like all my friends, they're just like, we just made music, they've always made music. I'm really interested in developing Rogen to the point where I can jam on stage with her, where she can train off my music and she can continue like say if I really do want to do this, like I don't want to slave away in front of the computer and make tracks like I am for the rest of my life. I don't want to do that, like I really don't want to do that. I would love to find ways to speed up the process but not take away any of the creativity and there are many things you do in the studio or to prepare plugins or you know that you can clean up and you couldn't help AI could help you do some of that admin or help you identify things that could be similar sounding that you could pull in without have to go through all the presets. I mean there are already things to do that. I mean splice I'm pretty sure I started to do that. I'd say that very few artists, I mean I know that is it Tim Beland that's like a lot on Suno and he's like very advocating and I understand it, it's fun, it's like it's super fun and it is hard not to not to go that I had a play, you know I have had a play and I was like well hold on a minute that's actually that's actually really cool, like I would actually might use that but you have to stop yourself because I do have, I do want to create standards that allow for musicians to be recognised for the work that they've done that these these streaming service that these AI models have learnt from and hoovered up but are not in any way caring at the moment seemingly about how to solve that problem and I really want to spend my time helping to solve that problem, helping to contribute to you know attribution for musicians of the past and the now so that they see themselves in these future models if they want to and if they don't find. One of the things that I've heard from AI companies before is that they foresee a time when pretty much everyone makes music. One person even said to me once that you know back in the day we could have never foreseen that people would go from you know if you wanted to get your family photographed you'd have to go to a photographer who is a professional and had the camera sit down with them at a special place but now we just take photos on our iPhone and this person was saying that they foresee a time when everyone just gets to make music it's not something reserved for musicians anymore I guess. Great. What's wrong with that? Yeah yeah I'm wondering what your thoughts are and like the the barrier to entry lowering so that everyone could feasibly do it. I really why not? So if you've spent many like 10,000 hours whatever of perfecting your craft you always have an edge um you know whether you're if you generate anything of these you know services right now you're just going to sound like 99.9999 percent of other things that do that too but if you have an edge if you have a real something there that connects with people and you use it differently and you go a little bit further out of your way and you create something that's genuinely standingly good then that's great that's really good that helps everyone move forward. I don't have a problem with that at all like I said that everyone takes way to our top where we can all put filters on it and clean out and the other thing is like why should music just be reserved for those who you know have had emphasis to like computers and programming and lessons if you need it and you know all the time or whatever money I don't know but I mean because initially um we all used to sing what we sang before we spoke um we we had songs that we would sing and we would dance and we could move our bodies and we felt that was just the way we expressed ourselves and now we're just scared to sing um so I think this is great you know if it helps us connect with our creative in ourselves um and allows us to feel more at peace with ourselves because there's something that's you know really connecting because it feels great to make me sing then I am definitely for it 100 percent I've got no issues with it if anything it's just going to what needs to happen which will happen in tandem with it is much better discovery tools to help people see through and hear through the noise because it is ever so much harder to find the good stuff out there as there are many many many issues well there's that there you don't really know if you're listening to AI Generative Music at the moment on on these platforms I have come across hundreds of generative versions of headlock and other songs that I've done that are definitely not human or largely not human creators as they come up quickly um and you look at the profiles and you try and find out and there's all kind of disappears and it's all like one label and nobody really knows and then it disappear and you know it's just like oh was it there at all um so there's definitely like weird stuff going on already and who knows who's benefiting from that you know the top level who knows um so it is just a mess and I think it's good because it has been a mess forever and now it's just really showing the problem one of the interesting things that I've been thinking about in terms of AI music is it certainly speeds up the process so we're going to get more music um especially I think we're going to see more music ending up on streaming services right now we already have 100,000 songs being added every day you know if we put AI aside we look at a few years ago when it was still a similar number of just a mass amount of music coming into the services every single day a lot of that music is from totally DIY artists I think the the revolution of DIY brought forth a ton of music but you know for the most part most people aren't listening to most of it some of it's not very well done and that's okay uh you know they they've released it it's a passion project whatever um but you know algorithms allow us to navigate through the 100,000 songs that are there on Spotify every day and so as AI comes in I think it'll just accelerate that process we'll definitely have even more songs um but this is not a new problem we already have a lot of songs to sort through so it just has to get smarter and better I think what we're going to be able to wrap into it is the story of the song so right now the algorithms are like pitch and tempo and tombra and like maybe lyrics or dynamic range or that kind of stuff that it's creating algorithms for us to go oh you might like this um but you might want different ways to search you might discover that you like you know a particular producer you know and if you're lucky you might I don't think you can search via producer at the moment on streaming services but you can search obviously through artists but you can't you've just about started to be able to search through composer um that you still have to be registered in a certain way not everyone's allowed to be a songwriter composer I've only I'm just allowed to be one even after how many 30 years I've been doing it um so there's it's not even at the basic stuff isn't sorted you can't search through drummer or um you know mixing engineer um but you might discover if you could that uh you actually actually the reason why you love these songs is because of the arranger or it's because of the particular way somebody pays a string instrument that in that um kind of inspires the midst of a song for someone yeah yeah and to kind of pivot the conversation a little bit you just recently celebrated the 20th anniversary of speak for yourself which is very exciting and I mean it's really crazy how many times these songs keep coming back up um I really enjoyed normal people um in the sync placement that a hide and seek had in there a couple years ago I've been on TikTok and I keep finding your work there as well I'm wondering why do you think this has continued to find new audiences organically over time I don't know maybe there's the nostalgia um maybe there's something comforting in it um I don't know I asked myself the same question I feel very lucky um I think I don't know maybe it's just a generation of people that 20 years ago you know were just starting out there with students or whatever and at that point for them it meant something because they heard it in the OC or whatever however they discovered it um and people at Ariana Grande they discovered it then and tailored that they all kind of discovered it then and now 20 years later you know they're massively famous pop stars and um they're just killing it all around the world um and they're introducing people as well um and the people that were watching those programs back then they're in cool jobs and doing different things and you know so they they chuck it in because it means something to them and I'm finding that not just in the music space in the like in the fan space but in a much smaller way also see it in the industry side I see the people that were my peers when I was younger in the industry we're also coming together again do you know what we're gonna we're doing things differently now and I'm finding I'm finding it much easier to navigate and to help shake the future of this industry because it's a different year groups coming through basically the old guard are going out and we are able to come together I'm like nearly 50 now I've already said and and it just it just feels like it's easier and then it will feel easier again for the next generation hopefully as they learn from our mistakes and see that as they point to be rigid and as they point to be possessive over things nothing fun comes from that just more pain yeah yeah earlier you were touching on like the music industry and things turning over and one of the things I find really interesting about speak for yourself is that it was first release I didn't realize this until very recently that it was released independently um which was probably not a very easy thing to do back then it's now become so much more popular to be an independent artist and to find success that way I'm wondering looking back on it um how do you think it compares being an independent artist now versus when you released that album 20 years ago um I feel like I was in a sweet spot because it was just as the industry was going um digital and streaming and there was iTunes had come out just literally a month or two before speak for yourself happened and there was only one front window there was like now there's like 50 revolving windows and covers and I've got lucky and I got chosen my album got chosen to be the album of the week on iTunes wow and I was like how did this happen you know I'm nobody I'm like I I just couldn't believe it that you I thought you'd have to pay for that spot but I did release the record myself I was independent and I ingested it myself I still have an account for that but even though I shouldn't um because I was one of the first and so I sent a message to the to iTunes uh like info out iTunes or whatever and I was like hello uh just a message to say thank you to whoever put my song my album on the front page I think it was maybe the UK iTunes and I was like thank you so much you don't know how much this means to me and I got a message back from this guy um called Denzel Feagelson um and he's he was the boss at the time uh and he was like oh well no one's ever thanked us would you like to come into the offices and I was like yes I went uh so I went into the uh I went into the offices I met Pete downturn who was like one of the first digital streaming um gave the music to the streaming companies like Spotify before they had their own way to do that and I went to there I was the first artist in the office so I have this like I was at a lucky point where there wasn't that much going on um you either with a label you weren't um and I and I reached this point where people were like wow an actual artist is coming in to the office this is weird it's not a music executive and it really worked it by favor um and it was much less scary to do independent than to go again with a major label it was just it just told me apart every time they just made a terrible job of of not promoting it or promoting it a good job of not creating it yeah because fru fru was on a major label right and that's uh then you were dropped after then you went independent my first album was called i make a phone my second album was fru fru with guy and then um fru fru they decided they didn't want to make another fru fru record and I was like fun um they're like we'd like to make an image in a heat record and I was like no thank you um you literally you couldn't have made a bigger mess of such a great record like we absolutely thought the details is one of the best things ever it was so um and now just so many other people think that but at the time they were like no no didn't get on to radio one there's not it's not gonna be as big as sugar baves and they didn't like give us any airtime tool or any any support and they've they've made us try and come up with some hits and put on the record it was just so awesome oh my gosh um so in the end I was like can you please just let me go because I just can't I just don't want to do this I don't want to do it and I think they thought fine um I figured then I'll just I'll give you first refusal which I didn't but they didn't care anyway uh they would just probably have to have it here and do me um and then it was just so much easier I basically re-morgaged my flat I took a hundred grand out because it had grown in value over it's this small flattened waterloo over a year and it was the only way I could get any money I did try the banks but didn't want to get any money so I re-morgaged my flat and then I I put all the studio equipment and I did it for a year and then I could have gone with a major label then but I was like I'm going to do it myself like I'm going to bring all the people that I've met from all the labels and all the times and I'm going to bring in my best crew and we're going to do it to get I'm going to pay them and I'm still going to own way more than I would have ever done if I'd signed to a record label um and I only just you know fru fru was released 25 years ago I think and I've only just seen royalties from the album wow only just we've only just recouped isn't that crazy that's crazy wow that is hard to believe I mean also that fru fru record is incredible and to think that it wasn't enough at the time for this label is is hard to believe um it's one that I've really enjoyed listening to to this day and I know a lot of other people keep you know finding it again I mean the streaming numbers are crazy I think it's always interesting to look at a streaming service in an album that was released before streaming and see how it does because some of them just don't fare that well and that's okay um but to have an album that is still being streamed that much um from what 25 years ago or so is really remarkable um but it's it's it's amazing now um in 2025 how many artists are waiting to sign longer they're getting better deals because they have the ability to be independent I'm curious like if you think you were coming up now do you think you would have also still chosen that independent path oh hundred percent yeah I really do not don't sign a major label deal like basically everyone who signs a major label deal after about a year they're like how do I get out of the steal how long have I got how many more albums do I have to do before I get out of the steal every artist knows that it's a stepping stone to becoming independent it's not it's not the place to end up you know it's just you got to make the best deal that you possibly can and they're always going to give you the worst deal and even if you think it might be good because of the way they were doing it and and also you don't know what they're doing especially in terms of AI training giving over licenses to what you we don't know what they're doing we don't know um it's not in I wasn't in I mean I'm not signed to send you anymore thank goodness but and they at the time I was I was actually like two years ago and they still had licenses of my stuff I was writing letters I was like do not in any event give my music to AI companies do not do not do not do this otherwise I'll take you to court and I think I might message that somewhere on a open letter I can't remember I didn't really get any replies um I thought it was very suspiciously quiet um but I did hear that backroom deals were being done at major labels with AI companies and that they were doing discussions and I I don't know what they've done but I'm glad to not be in it anymore yeah I mean I do think it's very interesting about uh record contracts as you sign it and um so they usually have some sort of a clause that leaves the door open for technology in the future that can't be foreseen at this moment um which I understand why they do it but it does leave yeah instances like this I mean who could have foreseen um generative AI in this form and that's something that I think about a lot with legacy artists or artists who are deceased um I've reported on some estates um offering the voice rights of a person who's been passed away for like 10 20 years uh to AI companies and I always wonder like if they could have even foreseen the generative AI would be a thing how would they feel about this use and unfortunately we'll never know um but yeah I think yeah it's tricky isn't it uh I guess they did okay um but uh yeah if it's the family then I'd they'd probably just be like yeah sure of course I'm dead do what you like yeah um that that would be my that would be my sounds one more thing I wanted to ask you before we start wrapping it up I I wanted to ask you about the gloves the mimu gloves I absolutely love those um there's something that Ariana Grande is used on stage quite a lot they're really wonderful to see in um everyone's different hands I guess um are you still iterating on that and improving it or do you feel like the product is done as is that we're still we're still isolating um we have we have a batch that's going to be delivered hopefully before the end of the year uh to like a few hundred people um we haven't yet managed to get them on the shelf um they're not you know they're very very good quality the best in the world uh for creating music and and there's they're expensive they're like two and a half grand a pair so you know it's like buying a it's like buying a computer but they are they're expensive because they're basically like tiny team we're doing our best to make them you know as best we can as cheap as we can but we're not like a mass you know we're not making millions of these things so it's still expensive and a lot of people that's obviously a big barrier um and it's a it's a leap to go to you know make me suggestively even though it feels telling you naturally do it um so we are missing out on the people who the impulse buys because it's like oh I want to do that I'm going to buy that yeah yeah and then the hit buy and then they're going to get it stay and then they're going to love it but then it's like oh even I'm not going to get it three months or maybe six months because there's takes on that and then they then buy yeah our problem really is that we need to have enough capital in advance to buy them to get them on the shelf and then I think we could see a lot we'll see a lot more people buying them because you can't you can't get credit for something if it doesn't exist um so once we can create a way they're on the shelf then people can pay over time over a couple of years or whatever it's much more it's too easy to stomach um but we're just we're a tiny company we haven't taken investment other than like friends and other musicians tiny bits um because we just want to do it our way you know um so we're hopeful that we can find a way to get it on the shelf um because I do think we haven't seen the end of it yet really even though it's been really hard we'd we'd have big chat the other day um but there's nothing else like them you know I know you can do it with the computer there will be you've seen videos of people using touch design or whatever and just kind of you know using the IR the camera and creating gestures and doing it in front of your computer I we know you can do that you can do that with us too with the clever software but there's nothing like being away from the computer being in the garden and making music with the trees uh you've got I've got a wireless microphone I've got my in-ear headset and I'm just out in the garden making music with whatever I want to make it with or I can go into the studio or I can go into the by the piano or in stage and it's just that freedom um of being able to conjure music in real time in real whatever space and not be in front of the camera. This has been such an interesting discussion I'm gonna end it off with some quick like rapid fire questions and then we'll do our playlist um I'm not gonna get rapid fire answers but I'll do my okay okay I just a couple words per answer and we'll see we'll see how it goes okay um not myself yeah who is an artist that you view as an innovator uh team eggsaw okay a song from speak for yourself that you look back on with great pride uh hide and seek yeah the coolest way that someone has paid homage to your work I don't know I really liked hearing like Kazoo version of hide and seek um I love that I mean no I really grand when she did a good night and go uh and she like I love it I love that she did that I like what I went on tour and I did my version of her version of my song um and no I really liked the way she did it she's so generous she really just has continuously shared the love uh and always supported me and she's just like enormous as she just has such integrity all the things she's been through I think she's so talented I really I know I'm going to work we are going to actually do something together one day one day we both said it it's just we've got to find the right thing at the right time so that we can really enjoy it and it's not just something like but it's gotta be but you know the same time it's like you don't want to build it up so much and then we never do it because it's like it's gonna be the perfect time um but no I really really want to work with every another I think um she's fucking great yeah she is what's the best film TV use of your song hide and seek I mean obviously the OC um because that was the first one but um I think the the most intriguing one was the saturday night live skit um because they didn't ask permission um they just they just did it and there quite a lot of people had seen the OC but way more people saw that skit and then well didn't even know where it came you know did they wait came from I suppose but didn't didn't know about the song it was like too many steps removed at that point and this was one it was one it ended up being one of the first means are on the internet of people just doing what to say and just all different like synths and version all all different versions of it um and I don't think it was an official synth version but people just you know edit stuff up um and it was amazing it was amazing to see but that was before monetization sadly on youtube so there were like billions and billions of uh bits of this song but I think it like goes under the radar of monetization because it's not quite 30 seconds but otherwise it'd be like I'd be like total million it and yeah that really is too bad that it was just a little early on the youtube thing but um I think that song still has a lot of legs online there are still people using it all the time um who is an artist who counts well actually you kind of answered this who'd be Ariana um what let me ask me the question um and artist who counts you as an inspiration that you enjoy listening to yourself well yeah Ariana um I don't know I mean I heard I like I really heard this person's music recently called Blarron not that I'm sorry a data cluster um and he was like totally excited so and I really like his music well I discovered lots of people that I'm like well I love this person's music and then I think like there's no way they're gonna know me because I'm just not cool enough and then like no I really have to be like I'm like they're guilty pleasure um which I find a little bit like how can I be a guilty pleasure I'm not cool enough um but I'm very lucky I I seem to find fans in all kinds of like unexpected places in like you know extreme heavy metal rock bands or noise artists or hip-hop or flagra app or you know just you know just like it's just it's mad how cross genre um the music gets uh really love that I feel I feel so lucky really to have yeah no people from all over and just feel like it's a possibility really hopefully to work with anyone if I if I really why do you maybe I couldn't I don't know um but there's so many people I want to work with I just I don't want to sit in front of the computer and write me the way that I have done over the last 30 years just like solitary mixing and just moving bits of audio around I have to find another way and I know that obviously are other ways that you can do modular you can you can work with a band um but I've just become a slave of my own technique really and that's why I really want to develop the live system with the Mimi gloves with generative AI of of motion but under an acceptable level of AI generative work so that I can conjure lyrics in real time with my own brain in real time um but also maybe with the the temperature or how many people in the room changes the way that the melody moves around or that the how fast the chords move depending on how closely after me or I just I want to create music in real time in a very different way than I ever have and that's really well spent my time on exciting well um okay so the final thing of our interview is the playlist part so we always ask everyone on on the record to help make me a playlist what is a song you can no longer gatekeep we'll just start there what's a song that you're loving right now that you want more people to know about I really do love uh this album by that one called sick elixir uh sick elixir I nearly had a car crash because I was driving and I've been waiting to hear it in the car so I could hear it not just online my head friends or in a wherever I wanted to hear it like loud and so I played it and I had had a car journey enough time and I oh my god I knew I just couldn't believe how good it was it just I was like I was swearing I was like if somebody looked in the car they would be like what the hell's that one doing I was like going oh my god this is even amazing I was just like I couldn't believe how good it was I was just laughing just I was like a gift being possessed um so the whole album is absolutely amazing but if I I can't really choose on because I've I liked all of them um but I don't know I like it there's one called dot dot dot NOS what is a favorite throwback of yours so a song over 10 years old when I was like 17 in canton when I was you know starting out and there was this band called acacia um and one half of it was guy six worth who obviously we'd get turned into being free free but the album that um that they made acacia um is so good and I think that one of the tracks that I love the most was called maddening shroud so I'd check that out it's from the 90s so long I'll have to check that one out and last but not least what is a guilty pleasure for you um well I just it feels a bit weird to say it because it's not really like a guilty pleasure but it kind of is I talked about this person earlier this this artist called data cluster and I found them on uh you know on this random playlist and I was I was my I was a little bit thinking that they might have been AI generative between some artists that I love so the guilty feeling is like but I love Max Cooper and I loved on Hopkins and they're like my I like know them um and there's such echoes of their music in this that I feel slightly uncomfortable around because it is so to the tee uh but the melodies and it's not it's not their music at all it's just data cluster um absolutely idolizes these artists and has studied them meticulously to the point where he just knows how to get those sounds I just would not have the patience to do that and I'm just like I'll just kind of make a sound like this um so I guess my guilty pleasure in a way is that because I would like to talk to John and and Max and see how they feel maybe they're like there doesn't sound anything like that how would you think that sounds like me all right oh sorry huh um so the one that I found on this playlist I'm going to try and find the name of it circa 13 it's had under a thousand plays on 45 and and he's got 899 monthly listeners but I was like this is just great and I really liked it and I was listening to it a lot and I couldn't touch with him I think he lives in LA might meet up with him and I go to NAM in January um but yeah data cluster and the track that I heard was called circa 13 but it comes off his debut album um I believe which is called coalescence awesome yeah okay I'll check it out it's really nice well thank you so much image in heat for coming to the show this has been so much fun I love talking AI with you and you'll have to come back soon thank you thank you my pleasure all right another big thank you to image in heat for coming on the show today and sharing her insights with us it's really good to get a pulse on how an artist is feeling about AI right now especially one who is still very hopeful about the future up next we're going to be diving into our charts round up where we'll highlight this week's biggest movers and shakers here is the top 10 of the hot 100 chart for the week of November 22nd rising to number 10 this week is back to friends by somber number nine is I got better by Morgan Wallen number eight is folded by Kailani number seven this week is mutt by Leon Thomas six this week is daisies by Justin Bieber opalite by Taylor Swift comes in at number five man I need by Olivia Dean rises to number four number three again this week is ordinary by Alex Warren golden by huntrix comes in at number two again this week and finally the number one song on the billboard hot 100 chart for the week of November 22nd for its sixth week at number one is the fate of ophelia by Taylor Swift thanks for tuning in to this week's episode of on the record and another special thank you to my guest image in heat for coming on the show and if you like today's show please consider hitting us with a rating follow a thumbs up all those things really help a show like ours to grow and to reach new people again I'm your host Kristen Robinson and tune in next week for another peek behind the curtain of the music business I'll see you then
Podcast Summary
Key Points:
AI is being increasingly used by artists in music creation.
Imogen Heap discusses the impact of generative AI on artists and creativity.
The importance of distinguishing between AI-generated works and human-made works is highlighted.
Summary:
The transcription discusses the rising trend of AI in music creation, with artists like Zania Monet using AI tools to produce songs that are climbing music charts. Imogen Heap, a Grammy-winning artist and technologist, shares insights on the impact of generative AI on creativity and the music industry. She mentions her AI voice model "AI-Mogen" and her work on oracles to protect artists' intellectual property in the age of AI.
The conversation delves into the need for identifying and distinguishing between AI-generated and human-made works to ensure fair treatment and recognition for artists. Imogen Heap also touches upon the evolving music landscape, where major labels may start signing AI talent and the potential for AI to push creative boundaries in music production. The dialogue reflects on the exciting possibilities and challenges posed by the integration of AI in music creation.
FAQs
Generative AI is changing the landscape of music creation by allowing for new possibilities and collaboration between human artists and AI tools.
Artists can use tools like oracles to manage their intellectual property and provide necessary context around their works.
There should be a way to distinguish between AI-generated works and human-made works to ensure fair treatment and recognition for artists.
Major record labels may sign AI-generated artists as the music industry evolves, but the nascent stage of AI music creation is currently offering exciting possibilities.
Collaboration between AI and human creators can lead to innovative music that pushes creative boundaries and explores new sounds and styles.
Artists can document and share their creative process involving AI tools to provide transparency and authenticity in their works.
Chat with AI
Loading...
Pro features
Go deeper with this episode
Unlock creator-grade tools that turn any transcript into show notes and subtitle files.