Dmitri talks with Chris Wieduwilt – aka the AI Musicpreneur – about his route from struggling musician to vanguard for AI empowerment in music. They discuss the variety and scope of AI tools that exist, touching on ethics, ownership debates, and the potential impact on human creativity.
Links to shoutouts mentioned in the episode, and the Music Tectonics Conference 2024!
AI generated music video talked about in ep
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!
Join the Music Tectonics team and top music innovators by the beach for the best music tech event of the year:
6th Annual Music Tectonics Conference October 22-24, 2024 Santa Monica, California
Episode Transcript
Machine transcribed
0:00:20 - Dmitri
Welcome back to Music Tectonics, where we go beneath the surface of music and tech. I'm your host, Dmitri Vietze. I'm also the founder and CEO of Rock Paper Scissors, the PR and marketing firm that specializes in music innovation, and I'm also the founder and director of the Music Tectonics Conference, taking place October 22nd to 24th 2024 in Santa Monica, california. Today, we're going to step into the future by stepping into the now. We're going to go deep on AI music tools, understanding what's out there, what they do and which ones you might want to check out. To do that, I'm bringing the AI musicpreneur onto the show, chris Wiedewald. From starting a garage band at 15 to touring the US and working with world-class producers, chris helped scale a startup from zero to 150k per month and worked with major brands like Lufthansa, goodyear, l'oreal and Porsche. Armed with this knowledge, he created the AI Musicpreneur to empower artists in the age of AI. Christopher seeks to provide the tools and knowledge he wishes he had when he was a struggling musician, helping artists navigate the ever-evolving landscape of the music industry. Chris's company offers a virtual assistant for music promotion, a comprehensive guide to AI music tools and a community of artists and industry professionals interested in AI music.
Chris, welcome to the show. Thank you very much for having me, Ddimitri. Yeah, I appreciate it. We've talked a lot about AI and music on the podcast and at the conference and we work with a lot of AI music companies in our PR work at Rock Paper Scissors, but I thought it would be cool to have you come on. You're posting a lot on LinkedIn and on your website about kind of the wide range of AI music, with a focus on what artists might be interested. So I wanted to use this opportunity to really get a little deeper, get a little more specific for our listeners. So I'm curious, chris when did you first get interested in music and AI?
0:02:13 - Chris
So my my grandpa my grandma actually was an opera singer before the second world war and I always tell her story when I explain how I got into music, just because the truth is she was this great person that loved music dearly and her heart was full of it and that made me want to be in front of an audience and sing to other people. So she kind of instilled that love for music. And how I got into artificial intelligence is I had this chat GPT moment, sort of that took me back to my 10 square feet room in east london where I was trying to make it a music over 10 years ago and I was struggling there massively and eventually had to move back home and give up my dream on a music career, but I I just wasn't able to make it sustainable.
so all I had was my following on myspace, basically, and when the platform sort of declined, you know, my following also declined. So when, when ChatGPT came around and Midjourney and all these different AI music tools, I thought, wait a minute. Major label artists they have teams of strategists, marketers and managers, but indie artists they basically have no team. And AI gives artists a legitimate chance to close the gap if they know how to use it to their advantage. And I said I'm going to be there helping them. So that's what I did I'm trying to show artists to get the tools I wish I had when I was trying to make it in music as well, nice.
0:03:39 - Dmitri
Got it Okay and as we broaden out before we get more specific, what's changed in AI that's made it the number one conversation topic in the music industry right now?
0:03:49 - Chris
So it's definitely the hottest topic in the music industry right now and, I would say, for good reason. Also, so there's all these new AI models like ChatGPT and DALI, that you can create human sounding text or images and audio and even entire songs. Create human sounding text or images and audio and even entire songs. You probably saw the Sunos and the Udyos of that, where you can just put in, basically, instructions and it makes music. So this is at the core of it. This is a big deal, because these AI systems, they can now mimic the styles and the sounds of real musicians at the end of the day, and they can analyze an artist's past work and then generate brand new music that sounds just like them. So with AI, anyone can create songs that seem to be made by their favorite artists essentially.
And that raises a lot of problems for the music industry, because if AI can easily copy the unique styles of human musicians, it could take away work from real artists. And it's not always clear who owns the rights to AI-generated music. Is it the AI company or is it the person who wrote the prompts or the original artist being copied? And a lot of record labels, musicians and policymakers they're all trying to figure it out. I would say how to deal with this new AI technology, and they need to set rules around using AI for music creation and make sure artists are still getting paid when their styles are replicated. It's a complex issue, but one that's crucial for the future of the music industry. So, in short, the progress of AI has really turned it into the biggest topic in the music world right now, and, as everyone really tries to understand its impact and implications, Right, yeah, and we'll get into some of the implications later in the conversation as well.
0:05:38 - Dmitri
But I appreciate that context. I mean, it seems like the technology has just gotten so far along that you know it's inevitable that it's having an impact on music, and it's really interesting to have somebody like you who's had that music, that musician creation hat, on now thinking about, well, what would I have done differently if I had those tools and how can I help others learn about those tools? So my intent with today's conversation is to catalog all these different ways that AI is influencing music today and for us to get a solid understanding of kind of like those main categories, the ways AI is being used in music. So let's break down those types of music, those types of music AI tools that are out there, and then maybe you could tell us what each one is. Maybe what are one or two companies or tools that allow you to accomplish each. I don't know. Do you want to just pick one out and start there?
0:06:27 - Chris
Yeah, let's go for the obvious ones, the AI music generators. These tools, they use basically AI to generate new musical compositions from scratch based on, for instance well, there's text prompts, text-to-music tools, but there's also other input parameters that you can set, like a genre, mood, instruments or examples, and some tools that are known in that space are Udio and Suno. A lot of you probably have read about it, but there's also other tools, like AudioCypher, for instance. That is a word-to-media music generator where you basically type in a phrase and it creates melodies and chord progressions that you can also use inside your DAW, and I could go on and on. There's Infinite Album, for instance, for gaming that you can use Live Score Music Loudly. Everyone's trying to kind of win the race right now loudly. Everyone's trying to kind of win the race right now. So it's interesting because it's very rapidly advancing at the moment.
0:07:32 - Dmitri
Yeah, it's interesting because a lot of them have different use cases. Like you said, some might be for video games, or some might work well in a DAW, some might work independently, which might work better for like kind of meme, social posts or for video production, music generation for video and so forth. But then also the user interface is different, right. So you've got some that start with dropdowns, right where you start selecting from either genres or feels, and then you've got others that are like text prompt. Are there other interfaces besides those two?
0:08:05 - Chris
And then you've got others that are like text prompt. Are there other interfaces besides those two? So let me say this, because you just raised that they're very, very different from an interface point of view and what I'm seeing right now is also from the UX and UI of those tools. So if you go into the tools, they're not necessarily. They don't feel like a musician tool always when you look at the terminology that is used. So I always take the analogy of when I tell my mom about how DAO works and she looks at it, it's basically like a spaceship to her. Set aside, she likes Star Trek, but she wouldn't know how to use it.
But if you look at suno and udo, they're, they're very, um, they're very tailored to not just musicians but content creators or for your casual birthday song that you want to, for instance, create, uh or any other occasions. And there's a lot of music that is currently created on Udio. They suspect around 850,000 songs a day that are currently on Udio. But the good thing about this is these are generations, but only a portion of those get published at the end of the day, in the minority, which I think are going to end up on DSP platforms in the long run. So, yeah, there's these text-to-music platforms and then there's ones with, like you said, predefined interfaces where you just like Soundful, for instance, where you just input the genre that you would like, the BPM, the key, yeah cool.
0:09:40 - Dmitri
Cool, so that's a big category. We could talk for hours and hours about AI music generators. What's another category we should tackle?
0:09:47 - Chris
Let's talk about AI voice cloning. It's really a conversation in itself. So basically, what it does is you can these tools that can clone a person's voice and generate new vocals or lyrics in that voice, based on text input, for instance and that can be useful for virtual singers and rappers, for instance. There are platforms like VoiceSwap, vocalist AI and also Kids AI Interesting Kids, for instance. They just launched, or they're about to launch, something called Kids Earn where basically they're paying artists on the platform that are using a certain voice model.
So to make sure that this is also ethically, because there's a lot of AI singing voice generators that have the Drake's and Ariana's on there where you can use them. So we have to get rid of those, definitely if there is no consent from the artists. So we have to get rid of those, definitely if there is no consent from the artists. But AI voice cloning is a great category because it can help vocalists and also voice actors monetize their voice. In the end, when there's a marketplace so everyone talks about, it's a threat for them and it's going to take away work. But I always think about it like this you cannot be at 10 plays at the same time and sometimes you have to turn down some projects that you would like to do but you couldn't do it because you're committing to something else, so that gives you the opportunity to be at more than one place at the same time.
0:11:15 - Dmitri
Sure, it's like you could license your songs for somebody else to record. You could license your recordings for people to play and, you know, entertain with, or you could license your voice to be used to create a new song. It's a new, it's kind of a new category of how you get your musical DNA out there as well. And that's cool about Kits AI because it has not only an ethical component but also like a monetization component for the creator side. I mean, it's funny the word creator starts to get more and more vague. You're like wait, who's the creator? Is it the person who's providing the data that the AI voice is getting trained on, or is it the person who's using it? It's kind of both. There are two different types, in a way, of creators and you know.
Some other interesting use cases for voice cloning are you could remove an accent if you want to sing in a different language. If you have a disability and you can't speak or sing, you could use it as well. And I've heard about some folks producers who are pitching songs and they don't want to use their own voice. They could use somebody else's voice to show, and actually some people are using it to pitch an artist in their own voice, to pitch a recording artist to say here I've got a song that would be perfect for you. This is what it would sound like if you sang it, which is pretty crazy. So it'll be interesting to see what other kind of use cases come out with that. All right, so we've done music generators, voice cloning what else comes to mind for you, Chris?
0:12:43 - Chris
We got AI mixing and mastering, so that's an art in itself, and what AI can do is automate the process of mixing and mastering audio tracks to almost like a professional level. So if you go on websites like platforms like Lander or Roex and like Master Channel, you can upload your song, you get a preview normally for free, before there's a paywall and then you can preview how your song would sound like with their mastering capabilities, and they're also working with some very prominent mastering engineers together. Where master channel, for instance, they work with wes clark, who's who's a grammy, I think, grammy award winning engineer who worked with a lot of great artists, and you can master a song in his mastering style, for instance, and then he gets paid every time that's done so. So that's also a great way for engineers. It's not completely going to wipe out the craft, but it's a great way for engineers to also monetize their skills.
0:13:45 - Dmitri
It's kind of like what we were talking about with voice cloning. You might have a voice. That's great. You can start to sort of leverage your IP by getting other people to create with it. The same could be true for engineers in this category as well, and it'll be interesting to see. Because the mastering yeah Lander came onto the scene.
Other companies have come on and done, have a mastering tool as well, and now we're starting to see, like Roax, these mixing tools, which will be interesting too, which I could see like. If you're a traditional recording artist that doesn't want to have somebody else generate your lyrics or your melodies or your harmonies or your voice, you still might say but actually I do hate hiring people or I do hate the labor intensive, intensive part. And even if you like, maybe you're ultimately going to hire somebody to do mixing or mastering, you might do like a some some like early, less expensive tests to sort of get it started, say, how do I want this to sound? Is it going to sound more like this or sound like more like that, and stuff like that. So that's pretty cool.
0:14:38 - Chris
Yeah, it's fun it's funny that it's. It's great that you're saying that actually, um, there's a lot of mixing mastering engineers that I talk to that say it's great that you're saying that actually, there's a lot of mixing and mastering engineers that I talk to that say that it's not always easy to work, for instance, with certain bands that have something in mind but they don't know how to get it across. And if they can send, for instance, a preview to that mastering engineer and say can you do it like that, but add there some more highs or whatever, or put more compression, that could be a nice workflow as well.
0:15:11 - Dmitri
Nice, cool. Okay, do we have some more categories of AI music that we should tackle here?
0:15:16 - Chris
We could go on for hours, dimitri. Let's do a few more. Okay, how about AI music videos? There's some AI music videos that came out in the last few months. There was one by a band called Washed Out that had I think it's called the Hardest Part which they used an AI music video generator called Sora, by OpenAI, who also do ChatGPT, and it's at a very experimental phase right now.
So, basically, what this does is the AI generates visuals and animations which can be synced to a music track to create music videos. So there's, for instance, a tool like Runway that you can use. So if you create an image in Midjourney, or you can use your own images that you created, took of yourself or your band, you can put them in there. You can animate them, you can make them more cinematic. There's a tool called NeuroFrames where you upload your music track and then, based on the track and the beat and the rhythm, it finds visuals that are fitting, but there's also something interesting. Fitting, but there's also something interesting. Uh, it's a new tool called muse and it's currently in the beta, where you essentially, you can create an ai avatar there and it's an animated avatar so that you can feature in your videos, so you upload 10 different reference images and then let the I bring it to life, basically and it's currently in the beta, but I had the pleasure to talk to the co-founder and CEO, lee Orr, and I feel like they're onto something here.
0:16:55 - Dmitri
That's cool.
So another tool that doesn't really compete with the music making side but can augment the work that artists are doing, to sort of add a visual element, to do some promotion and fan building with that visual side as well.
Give them some social posts, and I can imagine, you know, like just as MTV changed everything with making music videos like a popular form of entertainment, with social media it's hard to keep up with all the posts, but I can imagine, if these tools get good enough and artists get good enough at using them efficiently, they could have a music video every day for the same song you know. Or or yeah, come up with, you know, kind of other interesting visual elements that are pretty tough to do. I mean, you know, yeah sure, we've all got a phone in our pocket, we can shoot whatever, there's apps that help, et cetera, but at some point you know you gotta, you gotta decide where to put your time. So I can see this one being a real help on the kind of like promotion, fan engagement, fan building side of things. Absolutely. Are there any other tools or categories that we should be thinking about?
Let's do AI stem splitters oh yeah, we've heard a lot about stems and splitting. Tell us what's there and splitting.
0:18:09 - Chris
Tell us what's there. So, basically, what you do is uh, you can separate audio tracks into individual stems like vocals, drums, bass, uh, using ai, and then you can, for instance, remix the music or you can put a different sound on a synthesizer if you want. That. It was used very prominently on the last track of the beatles now and then, where the ai helped to isolate the singing voice of john lennon from the piano recording. That they had to isolate it and then remaster it so that it sounds like that he was actually in the room recording it in 2024.
And there's tools like Audioshake that are doing great work. Bandlab has some great tools. It's even for free. So if you go on BandLab, they have free stems with the tool, but there's some of them that are also directly integrated in your DAO, fruity Loops loops FL studio. They have one which is starting currently in one of their latest versions, also together with an AI mastering tool actually. So it becomes more and more integrated into your, your normal workflow. I would say that you have everything in one position cool.
0:19:26 - Dmitri
No, it's interesting because, similar to mastering, when it first comes out you're sort of like, ah, this isn't going to be as good or how am I going to use this? Necessarily, you start to see as the price comes down and the efficiency comes down for something like splitting out stems, you invent new use cases. It creates an opportunity where you're like oh, I never thought about doing that because it was impossible to do. Now that it's easy to do, relatively speaking, AI is doing the heavy lift. What else could I do with my music? What could I do with my demos? What could I do with remixes? What could I do with collaborations? What could I do with bringing people back from the dead? It gets kind of crazy.
0:20:02 - Chris
I think the whole sync licensing sector is also interesting in that regard. I think the whole sync licensing sector is also interesting in that regard where, when you're pitching your songs to an agency where you can essentially then provide not just one track but a track in different, a stripped-down version or an 80s version and you can do it quite easily with using AI to augment the process Gives you more opportunities for placements.
0:20:27 - Dmitri
There's one more that I saw you talking about online soundscape generators. What is that?
0:20:34 - Chris
So that's all, for if you find yourself in need to get that deep focus in, for instance, you could use a soundscape generator, so they basically generate ambient soundscapes, sound effects or audio environments that you can listen to. And if you want to have deep focus or you want to sleep better, for instance, you can use andel. You put it on and then, based on your sleeping rhythm, it kind of starts to get you into the rhythm of falling asleep a lot sooner than if you have problems sleeping.
0:21:12 - Dmitri
Right, they did that. Endel did that wind down collaboration with James Blake, which was like an eight or 10 hour thing that you put on before you go to bed or something and it gets you ready, gotcha. So is that less of an artist tool and more of a consumer tool, or are you saying artists should use it to wind down?
0:21:30 - Chris
I mean, of course, artists work a lot and there's a lot of things that they have to do, so it can definitely augment the process of, if they're in the content creation and strategy phase, that can be a nice addition to their workflow. But yeah, it's a very much consumer-oriented tool. But it can be used in therapy audio music therapy, right which is a bit more personalized. Usually, when you go to these consultations and they use music as part of the therapy, it's something which is predefined, something which is pre-recorded. But if the therapist can kind of make it more personalized to your taste and how you're wired, we're getting into the field of neuroscience then, kind of but that's an interesting case. But yeah, it's definitely also something for musicians if they want to do ambient soundscapes. Yeah, awesome.
0:22:26 - Dmitri
Well, you're accomplishing just what I wanted with this conversation, chris. Like we've very quickly kind of rapidly gone through a lot of the categories of AI and music and you've been so helpful at sort of just crystallizing what the categories are and giving examples and use cases and everything, but we're not done. I want to talk about marketing and AI and some other things too, but we have to done. I want to talk about marketing and AI and some other things too, but we have to take a quick break, so we'll do that when we come back.
0:22:53 - Shayli
Early bird tickets are on sale now, but they won't last long. We're organizing three amazing days in Santa Monica, california, october 22nd to 24th. Your conference ticket gets you into a keynote with Mark Mulligan of Media Research, high energy panels with music innovators, thinkers and builders, the Swimming with Narwhals, startup pitch competition, networking by the beach and more creative surprises to come. This is the best price for the best music tech event of the year. So get your ticket before early birds. Fly away Head over to musictectonicscom slash conference to get yours now.
0:23:37 - Dmitri
Okay, we are back, chris, and you helped us go through a lot of different categories of AI and music and helping people maybe get the whole landscape of what's going on here, but one that we haven't talked about is marketing. How is AI influencing music promotion these days?
0:23:56 - Chris
It's a huge one.
I think we can all agree that being an artist nowadays is harder than ever.
There's relentless hustle that a lot of have to go through to make it to the top, and everyone's trying to chase that viral moment. So where things like content creation AI-powered content creation comes in, it helps you relieve some of the stress of strategizing and writing content for your social media or video teasers, lyric videos, for instance, I built let's put an example, I built an AI system that creates a four-week song release strategy. So, basically, import your song info and it'll craft content, ideas that are supposed to resonate with fans, that are either educational, entertaining, thought-provoking and empathetic, and that takes the hustle out of creating content for a music release, because that's a huge issue that so many artists have nowadays and they're trying to get heard and be the loudest fight the algorithm, you know. So music artists. I want music artists to spend less on content and strategy and spend more time on actual music creation, which which they're struggling with. So if I can make the process and we can make that process a lot easier with AI, then that's definitely worth exploring how it can help artists.
0:25:17 - Dmitri
That sounds super cool, very cool. So I've come across some other uses of AI that I just want to bring out there, and you know a lot of them. Some of them may have existed pre-AI or pre-AI hype cycle and and and there's so. So some of this stuff is sort of like either they were using AI all along or AI is just augmenting what they do. But I just want to mention, you know, we've seen some interesting marketing automation tools. You mentioned one that you've built yourself.
There's, you know there's others that are related to programmatic advertising, pitching for playlists, etc. There's song discovery tools, which sure you know the big streaming services probably are already using AI for things around that, whether they're matching acoustically songs or in terms of social graph relationships of recommendation songs. But also a lot of tools out there seem to be helping large catalog companies, publishers and record labels, sync companies, figure out what's in their catalog or what's an appropriate song for a particular brief or pitch or something like that. We've heard about a lot of metadata cleanup and data matching tools. You know there's companies that do projections of royalties, future sales, streaming, monetization data or catalog valuation tools.
I'm sure a bunch of them are using AI, whether they say it or not. Maybe they don't use AI but they say they do. Maybe they don't say they use AI but they actually do. And then there's other fintech tools, I think, to help specifically artists and labels and publishers find and collect revenue, and it could be a wide variety, whether it's on the composition side or the recording side, but a wide variety of different royalties. I don't know that we need to go through all of those categories, but just kind of like a nod to the fact that you're very focused on supporting artists, and but just kind of like a nod to the fact that you're very focused on supporting artists and so you've given some great kind of summary of what are some artist tools. Is there anything else in those categories that I mentioned that you want to talk about or anything missing from that?
0:27:16 - Chris
You touched on a lot of different things actually, so it sounds very, very good. One thing that just came to my mind actually is I think it's very underrated it which is like targeted audience discovery. So something that I've been playing with in the last few few weeks is a tool called perplexity, ai. So it's. It's basically. It basically works like if you're googling a question, but it's using ai to combine different sources of information and give you a more personalized and tailored answer. So how about, for instance, you use something like perplexity, you input your streaming data that companies have, labels have access to for their artists, and you can create a so-called prototype persona of your ideal fan, where you know about personality, where they like to go to shopping, like for records, habits, interests, anything. Because what that does is it gives you more insights into doing personalized strategies to actually reach the right audience on social media platforms, if you know what your fan likes, if you know what they don't like. So that makes this highly targeted promotional campaign to reach the right audience.
0:28:28 - Dmitri
Wow, I love the way you're thinking about this and digging into all these tools and figuring out different ways to kind of like find some efficiencies on some stuff that's really actually very hard to do manually as well. So that's super cool. So this has been great. I want to kind of like widen out a little bit more. Since you speak to artists, I'm curious how you think about where things are heading.
How do you address artists who are convinced that AI is coming for them? I mean, I think you've shown some great use cases that have nothing to do with the creative side of music creation. That could actually help, but I think there's still a tendency to sort of be concerned like, oh, the more I depend on AI, the more jobs are going to get taken away. The more everyone depends on AI and they start posting their own music, the more crowded the market's going to be. What's going to happen with human creativity? I went to school for this, or I trained for this, or I did my 10,000 hours at shitty bars all over the world so that I could be an artist, and now some tech guy is going to come in and program things to replace me. How do you talk to artists about this?
0:29:28 - Chris
So one thing that I like to do is I feel like we need to understand history first, so to sort of like understand why artists are currently on the fence, and it's usually the fear of the unknown. I would call it. Ai is different, but history is sort of like repeating itself, in the way that throughout music in history we've had multiple new inventions that kind of pushed the envelope. If we remember the old phonograph, for instance, that was a real game changer back in the day, which let people record and play back audio for the first time, and but it also sparked a lot of major fears of folks that were worried it would make live musicians obsolete and that artists would lose control over their works and not get properly compensated. How about electric elements? I mean, they totally opened up new musical frontiers as well, like rock and roll. But you know they rubbed some purists the wrong way. Of course. They thought that electric sounds just sounded unnatural. Plus there were, you know, worries that these new electric instruments would replace traditional ones. Same thing with internet and streaming. That came along and revolutionized our music distribution. But with that came also also a lot of concerns about piracy and unsustainably low payouts to artists. So lots of disruption, for sure, but there's also a ton of creative potential.
So most artists that I talk to they're actually very curious about AI and if they're fearful, I tell them try to fight fear with knowledge. So those that are convinced that it's coming for them mostly have never experimented with AI before. They don't know what it can do for them and they're, like I said, they're sort of like afraid of the unknown, and that's completely understandable, I think. So here's how I usually like to approach talking with artists. I ask them about their daily routine of creating music and promoting, and we try to identify pain points together. The the truth is, I mean, you don't need ai to win fans to create music, but the reality is, ai makes you better and you can't do it all at a high level where, if you're serious about building a career music, you know you got songwriting and cover design and content creation and video production and mixing and mastering. I mean that's that's already like five to six people. So a lot of artists are asking me what's the secret to success with ai and I?
I give them a very, very simple three-step framework that I use and teach to artists, which is step one identify your pain points. Let's say, for instance um hey, my lead singer goes on vacation for three months, but we still want to continue recording and explore new ideas. How can we manage that? Step two try to formulate a measurable goal. I want to still record music even in the absence of our singer. And then step three would be let's think about things, how to optimize your workflow. So how about we clone the voice of your singer? You use it in recordings to prototype songs and once he's back, you can just add his part to the final recordings oh, wow, that's super cool.
0:32:37 - Dmitri
I love the uh, the specificity there and also just the sort of systematic approach to it. Um, so you're saying identify the problem, come up with a workflow and then, like, implement it. Basically, yeah, awesome, love it. Do you have any concerns about the human creativity element getting lost in all this, the more we lean on generative automation tools for composing and making music?
0:33:02 - Chris
I think I think these generative tools like ChatGPT, claude, and also AI music generators, they do offer a lot of exciting and creative possibilities, but there's definitely concerns about over-reliance that I have, which can lead to a decline in genuine human creativity and self-expression.
When human creativity and self-expression because, at the end of the day, I think true creativity that stems from the experiences that we go through, like the emotions and the perspectives that we gain right, and those are like elements that AI currently lacks, as it can only mimic the based on existing human creative works, basically.
So using AI as an easy way could hurt the variety and experimentation that help art and creativity grow and improve, and that could definitely lead to creative work starting to look and sound the same. However, I think AI can be a very complementary tool which gives you new ideas, exposing creators to a lot of different possibilities they may have not considered on their own. So I think the key is finding the right balance at the end of the day and using AI thoughtfully, without thinking of it as a full replacement for the passion and the hard work and the personal expression that, at the end of the day, makes great art right. The most powerful and meaningful art will probably come from the collaboration of human creativity and AI's ability, not just from one or the other. I think it's about leveraging the strength of both to push the boundaries of what's possible, while we we try to preserve the essence of a true human expression.
0:34:54 - Dmitri
I mean you could flip it on its head too and say, well, because so many other people are using AI and things are going to start to sound more derivative, more similar and there's less innovation. It gives artists an opportunity to figure out well, what makes me not sound like AI, or derivative content too, which doesn't mean you can't use AI to do that. It just means you have to have some creative human or natural inputs that go along with it as well. I always go back to this concept from the writer and researcher Clayton Christensen, who wrote the Innovator's Dilemma, who says there's nothing new to be created on the earth. You know, like all the carbon that makes up all organic matter is not, doesn't you know? It's all already here up? All organic matter is is not, doesn't you know it's all already here. Any kind of innovation humans do have to come from taking, uh, existing pieces and putting them together, and that's the. That's the thing about innovation is it's taking seemingly unrelated things and putting them together. You mixed your chocolate with my peanut butter and all of a sudden, you have a new, a new, something right, um, and I think that's true with music too, and can be true with AI. So I like what you're saying about the possibility of new creativity and also this idea of turn it inside out and say well, actually, if everybody else is doing derivative stuff, how do I insert something organic, something human, something never thought of in this way before, to integrate with these other elements too? So I like the balance you take, chris. That's super interesting.
But there are legal and ethical sides and I'm curious to talk a little bit about that as well. You brought it up at the beginning of the conversation, so I know you're obviously thinking about it and I think of the discussion kind of in two buckets. There's sort of the scraping and training of the data. There's some large tech companies that are scraping music from everywhere without a license and creating this pathway for millions of songs to be created without any attribution or payment to the original composers or performers of the music that they're training their AI on. So that's one bucket scraping and training. The other one I think about is the ownership of the output, and you talked a little bit about this earlier too these implications for who owns a song that's created in partnership with a technology company and that having a strong influence on the creative output of the finished product. You know it's like a collab between this artist and this software programmer or something, or this AI company or whatever. How do you think about these legal and ethical sides of AI, Chris?
0:37:21 - Chris
It's definitely a major ethical concern. So if we look at and there's other movers in this space that have looked at this, for instance, ed Newton Rex, who has written numerous articles for Billboard and other publications, where he inspected Suno and audio and was sort of like testing it against copyrighted material to see what it would come out of so it there's definitely several issues here with copyright infringement, lack of attribution and data privacy and undermining creators. At the end of the day, maybe they don't want to be in that training set that inspires someone else to create something from their work which was very personal to them, but I think we're starting to move in the right direction. It's always legislation is too slow and tech is too fast, but there's companies like Soundful, for instance, so they use session musicians and they record their own material and the output. It might not be as good as Sunod Audio, who are suspected of using the copyright material, but it puts the human at the center and compensates them. Generally, I think it's fine to use if training data includes copyrighted material, but it needs to be transparent and rights holders need to be compensated for that, and there's a huge chance also for songwriters to get paid at the end of the day, in the same way that they can use the AI voice clone and sell it on a marketplace. So I think there needs to be a framework for properly licensing and compensating rights holders when their works are used for commercial AI training. And I mean we're seeing a lot of initiatives like the Elvis Act in Tennessee and the proposed well, they actually went through the EU AI Act. Those are all steps in the right direction.
The second thing that you're talking about ownership of output right, all right, let's talk about this. It's a tricky issue of who owns the rights to the music that's created using AI. It's really a complex and a murky area, I would say, to be honest. If you take the whole question of human ai collaboration, when an artist works together with ai to system to make music, there's a lot of debate around whether the human should retain full ownership of it. If the ai company that created the system has some stake in it too, it's it's not exactly clear out at the moment. And there's issue also.
You touched on derivative works. If the AI is generating new music by training on existing copyrighted stuff, are those new tracks considered derivative works that need to be licensed? So that's a big legal gray area and there's some platforms like Stable Audio, for instance. So if you want to claim full ownership over the music and use it that the ai generates, it's it's a work for hire, since they take the creating. But that's this, it's.
It's a super controversial stance, so you have to. You can create the music on there for free, but you only get a personal license. If you have a subscription, then you get a commercial license, which is a good way to to say. But the bottom line at the end of the day is we really need updated copyright laws and licensing models that just work for everyone, that clearly define the ownership rights when AI is involved in that creative process. So something like that Grimes, for instance, did I don't know if you saw her Elf elf tech platform, but she uploaded her voice clone and then everyone could use it and the ones that you really liked. She did a a split copyright deal of 50 50. Put it on the dsp. If you go on on spotify and you search for ai grimes, you will find just the ai generated tracks. So I think the the the challenge is just that the law is slower right now than the rapid advancement of AI.
0:41:20 - Dmitri
I want to squeeze a little more out of this, but we have to take one more break. I have some more questions about the ethical and legal implications, and we'll get to that when we come back. The news cycle of the music industry, and innovation in particular, is accelerating at such a fast pace it can be hard to keep up. That's why I launched Rock Paper Scanner, a free newsletter you can get in your inbox every Friday morning. Check out bitly slash rpscanner. That's bitly slash rpscanner. I scan hundreds of outlets for you, from the music trades to the tech blogs, from the music gear mags to lifestyle outlets. So that you don't have to, I handpick everything music tech, including industry revenue numbers, ai, cool new user tools, the live music and recording landscapes, partnerships and acquisitions and everything else. A Music Tectonics podcast listener would want to know. Open a browser right now and punch in bitly slash rpscanner to sign up right now. Go ahead, hit pause and go to bitlee slash rpscanner, or find the episode's blog post on musictectonicscom and find that link. Happy scanning, but for now, happy listening.
Chris, this has been awesome. I feel like, in addition to cataloging everything, you really are thinking about all these different implications of the. You know what this means for artists, what this means for society and so forth, and you started talking a little bit about different laws and legislation. That's probably going to be the solution for some of the ethical dilemmas we talked about in terms of attribution and monetization on the scraping and training side, as well as in terms of ownership of output on the outside of that too, and there's been several AI ethical campaigns that I've seen unite. You mentioned Ed Newton Rex. He has a nonprofit now called Fairly Trained, with the goal to certify what companies are getting consent for their training data. I guess that's pre-legislation. At least it could have like a fairly traded data kind of checkmark or something.
The Human Artistry Campaign has emerged, which appears to be more of a manifesto that companies can join as members to show their support. I remember our PR client, bandlab, signed onto that very early, and then another one called AI for Music, which is a similar principles-based publication initiated by Roland, the musical instrument company, and UMG, the label, and with supporters from a lot of the musical instrument and music-making software industry. Kind of a slightly different bent there. And so there's these sort of like voluntary or kind of ethical stances, these manifestos that people are signing up to, and then you talked a little bit about the kind of some of the laws that are getting passed both in the United States and in Europe. I'm curious are you expecting large lawsuits for companies that haven't secured consent for training data? Do you think that's going to be some of the pathway for how this gets decided, or do you think it's going to happen more proactively? Do you think any of these sort of movement concept ideas are going to be the way to go? What do you think is going to happen?
0:44:21 - Chris
I think everything that is happening right now these ethical ai campaigns that you mentioned fairly trained and human artistry campaign uh, they're. They're aiming to establish industry standards and that's great, to protect artist rights, and you, we see governments that are also stepping in. But, uh, while getting the music industry to agree on clear ethical AI rules, it'll be challenging, but there's definitely reasons to be optimistic, I think, as more companies face these legal risks or public backlashes for misusing AI. It's just going to be pressure that is amounting to use it responsibly. At the end of the day. There's companies that are getting lawsuits now for using Drake or Ariana or the Whitney Houston's now and they're removing it from their platforms because they're afraid of the backlash and what that's going to mean for them and their business models.
So there's things that are happening in that and the path forward, I feel like, is involving everyone. I mean it's about collaboration. Right, we need the artists in the room, the tech firms, the industry bodies and the regulators. They need to work together. When Suno, for instance, they raised $125 million at a $500 million evaluation just the other day, and they're working with microsoft to bring to democratize music creation, but one of the things that they also said in their announcement was they want to be very close to music artists. So one of the artists that they're closely working with as an advisor is free low, for instance, or plow.
He's also the founder of royalio and I think that's a step in the right direction. If you want to create a tool that makes sure that everyone is being heard and no one is violated, then it's, I think, essential to have someone, or at least a few, in the room to address and raise these questions, because if you're just a developer let's say you're just a developer and you're coding away you might not think about the implications this might have on livelihoods. So that ethical voice in the room is going to make a lot different and, I think, ethically wise, we're going to move in the direction where we just get more clear guidelines of what we can do and what we can't do.
0:46:43 - Dmitri
Amazing, wow, super helpful to kind of get your perspective on that, because I think it's really important and it's going to shift how things emerge, how companies launch and how people engage with these platforms that we've been talking about this last 45 minutes, so let's wrap it up. Where do you think things are going to end up with AI and music in the next three years? What are things going to look like? Is this going to get mass adoption? Is it going to continue to have kind of mixed reception? How's music going to change as a result of all this, let's say, over the next three years?
0:47:15 - Chris
I think over the next few years we'll definitely see big changes as ai, you know, becomes more common also in the music industry, and but this progress that will definitely come with a lot of challenges as people try to figure out how the right way to use this new technology. So in music creation I, I think ai tools will get a lot better and more available, allowing artists to work with AI in new and creative ways, and we're already seeing that with new tools being directly integrated in the DAOs and optimizing the workflow. But there will, like we touched on, there will definitely be a lot of issues with lawsuits and public backlash against the companies that are using copyrighted material. We just need to find incentive structures to also circumvent this kind of behavior. So if something like the I think it's the there's a new bill which is proposed at the moment if that one goes through and goes through Congress, then that would allow lawmakers to. Then that would allow lawmakers to. Before you launch something, you have to have a thorough audit of your training data, what you're using and then, throughout the process of you launching your platform and your algorithm that is powering the platform, they need to be monitoring audits to make sure that not just for the first audit. You you were, I don't know, removing some of the data and then you launched and then you put it back in. So, yeah, the industry you know they may not need to come up with new licensing agreements and standards to deal with this. I think content creation and music promotion will be heavily driven by AI. The kids that are growing up right now with ChatGPT I don't know if you have kids or like in your immediate environment that are already using ChatGPT I feel like it'll be second nature to them to use these tools. In the same way, like with AI music creation tools, I'm hoping that people will still be learning instruments and how to play an actual physical instruments, but they use AI to augment the process and come up with new ideas.
There's another interesting field which I think has a huge use case, which is live performances. So I don't know if you saw the new Elvis evolution show in London, but it's kind of blowing my mind with how they're using AI. So they basically taught the system everything about Elvis his voice, his moves, his whole vibe so that the result is this hologram that's uncanny, lifelike, but the coolest part about it is that the AI-powered Elvis can actually interact with the audience in real time. So it's very different. It makes that experience feel a bit more immersive, and they've nailed all the details to the sights and the sounds and even the smells of an Elvis concert. So it's bringing the king back to real life in a really unique way, which is cool. That is wild.
But yeah, there's other avenues. Ai will also get better at recommending music that people will enjoy, help listeners discover new artists. Um, and, at the end of the day, I think, there's these thorny issues that need to be worked out, and it's going to take a lot of like as a collaboration between all the key players artists, the tech companies, the industry groups and the industry groups and the government. Everyone's going to have to come together, I think, and find that balance between fostering innovation and making sure people's rights are protected. So it's not going to be easy, I feel like, but it's crucial if you want to harness the full potential of what AI can do in music in a responsible way of what AI can do in music in a responsible way.
0:51:02 - Dmitri
It's so nice, in an era where our political leaders are all over the place and saying crazy stuff, to have somebody like you come on the podcast with such clarity and balance. Chris, chris Wiedewald, thank you so much. You're the AI musicpreneur, aimusicpreneurcom. Search for him on LinkedIn via his company name and, if you're looking for him, those VitaVault. Those are W's for the American listeners there. Chris, this has been great. Thanks so much for joining me. Thank you very much, dimitri. Thanks for listening to Music Tectonics.
If you like what you hear, please subscribe on your favorite podcast app. We have new episodes for you every week. Did you know we do free monthly online events that you, our lovely podcast listeners, can join? Find out more at musictectonics.com and, while you're there, look for the latest about our annual conference and sign up for our newsletter to get updates. Everything we Do explores the seismic shifts that shake up music and technology, the way the Earth's tectonic plates cause quakes and make mountains. Connect with Music Tectonics on Twitter, instagram and LinkedIn. That's my favorite platform. Connect with me, Dmitri Vietze, if you can spell it, we'll be back again next week, if not sooner.
Let us know what you think! Tweet @MusicTectonics, find us on LinkedIn, Facebook and Instagram, or connect with podcast host Dmitri Vietze on LinkedIn, Twitter, and Facebook.
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.
留言