Skip to main content

SC-044 · Live · WISE

Exploring AI Ethics and Music Innovation

Guest: Dr. Martin Clancy, Founder of AI-OK and member of the IEEE Global AI Ethics Committee

Summary

Dr. Martin Clancy is an audio engineer, researcher and member of the IEEE Global AI Ethics Committee who created AI-OK, a certification concept for generative AI music. He argues the industry cannot regulate as fast as AI moves, so AI-OK works less like a fixed rulebook and more like a signal that flags something once it crosses a line nobody foresaw. The label is revocable, and Clancy built it with people who also helped create DDEX, the music industry's own metadata standards agency.

Clancy traces today's fight over AI and copyright back to a legal loophole originally meant to protect the film industry, which let platforms such as Google and YouTube host user uploads without full licensing responsibility, so YouTube spent years operating effectively like an unlicensed jukebox before licensing caught up. He sees generative AI doing something similar at far greater speed, letting anyone create finished music without tools they have spent many years building themselves.

Listeners currently say they prefer music they are told is made by a human, though Clancy expects that preference could shift as tastes change and AI-made music becomes more common. His longer view is that once powerful tools are available to everyone, quality of thought and intent will matter far more than which tool made the sound. He is now talking with senior AI and record label executives in America and hopes a philanthropic backer could help fund AI-OK's development.

As of the episode's release on 25 April 2024.

Key takeaways

  1. 01Clancy created AI-OK, a revocable certification label for generative AI music, because he believes AI moves faster than regulation can follow and needs an ongoing signal rather than a one-time rule.
  2. 02He built AI-OK with people who helped create DDEX, the music industry's own metadata standards agency, rather than starting its technical side from zero.
  3. 03Clancy traces today's AI copyright fight back to a legal loophole meant for film, which let YouTube host user uploads for years while functioning like an unlicensed jukebox.
  4. 04Listeners currently say they prefer music they know is made by a human, but Clancy expects that preference could shift as AI-made music becomes more common.
  5. 05He believes the tool someone uses to make music will matter less over time than the quality of their thought and intent, since powerful tools will be available to everyone.
  6. 06Clancy is talking with senior AI and record label executives in America and hopes a philanthropic backer might help fund AI-OK's development.

Chapters

  1. Why AI ethics can't wait
  2. Martin Clancy's path into music
  3. Breaking down where music makes money
  4. What AI-OK actually certifies
  5. The loophole that built YouTube
  6. Defining AGI and its risks
  7. Why intent will matter most

Guest

Questions this episode answers

What is AI-OK and what does it certify in AI-made music?

AI-OK is Clancy's certification concept for generative AI music, meant to work like an ongoing signal rather than a fixed rulebook. It flags a product once it does something nobody foresaw and had no way of preparing for, and the label can be revoked if that use later turns out not to be okay.

Why does Clancy say we can't just regulate generative AI music?

Clancy argues AI is developing faster than lawmakers can realistically respond to, so writing rules for today's tools does not prepare anyone for the next wave of AI behavior. That is why he built AI-OK as a continuing signal instead of a one-time law.

How did YouTube get away with hosting unlicensed music for years?

Clancy points to a legal loophole originally written to protect the film industry, which platforms like Google and YouTube used to host user-uploaded content without taking on full licensing responsibility. As a result, YouTube functioned for years as the world's biggest music jukebox while effectively operating like a pirate radio station.

Do listeners prefer music made by humans over AI?

Right now, yes, according to Clancy: audiences say they prefer music they are told was made by a human. He adds that this is a current fashion rather than a fixed law, and expects tastes could shift toward accepting or even preferring AI-made music over time.

Will the tool an artist uses to make music matter in the future?

Clancy thinks it will matter less and less, because powerful AI tools will become available to almost everyone. What will separate artists, in his view, is the quality of their thought, their intent and having something real to say, well beyond which specific tool produced the final sound.

AI okay is kind of like the bat signal. And so it doesn't matter whether it's the riddler or whether it's the joker
Dr. Martin Clancy

Episode notes

Learn about the intersection of AI, ethics, and music innovation with Dr. Martin Clancy, a leading expert and member of the IEEE Global AI Ethics Committee. This episode delves into how AI is reshaping the music industry, from creative possibilities to ethical dilemmas and financial implications. Dr. Clancy shares his journey in music, the foundation of his groundbreaking research, and his vision for the future of music creation in an increasingly AI-driven world.

Highlights:

  •  The evolution of AI ethics in the music industry.
  •  Dr. Martin Clancy's insights on generative AI and its implications.
  •  The financial landscape of the music industry amidst AI advancements.
  •  Philosophical and practical discussions on AI ethics.


Topics

Transcript

Transcribed from the recording by the production team. Names and terms may be misspelled. Every line is timestamped: select a time to play from there.

Read the full transcript

The speed of AI development doesn't suggest that we will have time to legislate these changes, these emergent behaviors as they appear. And so for me, what would not be okay is something that we didn't foresee and that we had no way of responding to. So on a very simple high level, but also actually this is pretty much it, AI okay is kind of like the bat signal. And so you go, It doesn't matter whether it's the riddler or whether it's the joker or you know, you just dig it up and you go, okay, this is not cool.

What makes some products with Genotype AI okay? And what makes other products with Genotype AI not okay? What is okay? Martin Clancy is an expert in this field with AI okay. He's doing one of the most important things in the music industry. And just three years ago, no one understood him. Hey guys, and welcome back to Sound Connections podcast. I've been looking forward to this episode. It is something I think a lot about. It is AI ethics with Martin.

Fancy, welcome. Hello. Good morning, Jacob. We met some time ago at a conference that was really interesting. And I first heard about the concept of AI -OK. And it is one of the most relevant things to talk about in this day and age in the music industry. It's about AI ethics. And we're going to deep dive into a somewhat philosophical and practical talk about this whole area.

As some listeners may know, I'm very much involved in AI tools and using that in real life, but I consistently have conversations with my colleagues about what does this mean? How should we act? What should we think? So Mati, thank you so much for being here today and helping us navigate that. For the sake of the audience, could you briefly describe who you are and what you do? Oh, well, I'm an Irishman and I've worked in music for the last 40 years.

´I'm doing a variety of jobs, pretty much any job you can think of at some stage or other, I've done. And that is the basis of my research. I went back to school, I left school early, had a band that got signed and went off touring and went back to school.

about 10 years ago, and did a PhD about stuff that I saw happening with my students, and was curious about, and that was the ethical and financial implications of a thought experiment of what would it be like, and this is a crazy idea, if machines could make music that was indistinguishable from that made by humans. That was the, I remember at the time everybody was kind of saying, look, I don't really think this is a question, and I was thinking, well, maybe it's not, but, and...

I think it is now. Yeah, it couldn't be more timely, especially when things have become so easily accessible with audio and suino sort of launching. Of course, they are just like one product out of many, but they have made the thought experiment public at least to have a wide audience. Mati, before we start talking about this extremely interesting topic, could you just help me run through, you briefly mentioned it.

What is your story to the music industry and where you are today? You started 40 years ago. Let's start with that. Yeah. So I was in school and I was a post -punk, that kind of era of DIY musicians. And the first recording I ever made was with a singer called Sinead O 'Connor. She was about six months younger than me. And accidentally it kind of took off and I formed a band and the band got signed and we signed to U2's label originally and then to Ireland Records and then to Virgin Records and

some success. Not enough that the world will tease Kel for anything so I can't complain about diminishing returns. And we imploded and that's how I learnt about the music industry, the kind of hard way about in terms by actually having to litigate. recording contracts and understand that area. So I did that and then I decided to work independently in the independent sector and I worked as a music manager for a number of years successfully and left music in the around 2000, 2005 and moved into other areas because I was fed up with music and fed up with the business. But I got back into it by accident. My late wife started sending recordings I was making to a

free festival in America. And I ended up having some dance hits accidentally in my 40s, which was surprising. And I can talk to you a little bit more about that. And I was approached by a company called Ableton, who I was speaking about in interviews. And I became a certified Ableton trainer, of which there are, I think, 200 of us in the world.

And so I started teaching a new generation of musicians these kind of peculiar DIY techniques I had. And that led me to AI because I noticed that my students, when I introduced tools that were marketed, marked AI, behave differently. And I was curious about that. And that very simple, what does that mean? Brought me into this much longer research program and intervention.

approach that I've now kind of ended up dedicating my life to. That's how we are here today. Does that make any sense? Oh, that was perfect. Perfect intro. It sets the scene. I'm curious to hear about also the fed up of the music industry because that's, you know, the only person in life that said that. But let's just sort of venture into the core topic today. AI ethics, its implications also on the financial aspect of the music industry. You started.

10 years ago, people didn't really understand why we're doing this. Now everyone understands why we're doing this. You've done a PhD in topic. You've also released a book relating to this. Can you just tell me a bit about the book? What is that focusing on? Yeah, well the book, effectively the book became an extension of my PhD research, which is still the basis of everything that I do. And that's very simple to understand. I looked at what's going on. So who are the...

actors who are in this AI field. What does it mean legally? And so I explored the implications that primarily resulted in music copyright as far as I was concerned. I felt there were many, many holes inside of music copyright to the challenges presented by these new technologies. And that led me to go, well, what do we do? And that brought me to ethics. And as a result, I worked with the IEEE Global AI Ethics Committee, which I ran for three or four years trying to look at this with other people who were in the general AI ethics space and then I applied AI ethics to music. And then at the end of that, I thought, well,

let's see if we can think of a solution. So in relation to your question about the book, what I did was I went, well, we call this transdisciplinary research where you have to become an expert in philosophy and law and ethics and everything else and computational creativity. So I just wrote to a number of people and said, hey, this is who I am. I'm a student. I'm writing about your work. Do you have any thoughts on this? I'd like to do a book that shows where we all are on this. And I was very lucky that a number of people, from David Cope, who was one of the originators of AI music, I don't hear much about him. And it's really strange that we don't hear much about him. To Scott Cohen, who I know you're interviewing later on, who was very generous.

to Jacques Attali, who had written about this in 1977, Holly Hardin, Miller Parchard, who created the Max Sun PD programming languages, and they all just wrote chapters, or allow me to interview them, and that became the book. And we've just, and this is a heads up for Scott, if you're interviewing him later on, we've been recommissioned to do a new version where everybody gets the opportunity to reflect.

on what they thought was happening back in, I think we completed it in 20. 2020, so it came out, or 2021, it came out in 2022. So it's gonna be very interesting for everybody to kind of look back and go, you know, in Scott's case, he got everything right. In fairness, it's actually very interesting. Normally, this would be something that you'd feel maybe a retrospective after 20 years, but it's indicative of the changes that have occurred in the short space of time that.

in two years, we're looking back to something that's almost generational in terms of change. I'm going to ask you something controversial, Martin, because the listeners who listen to this podcast are innovators in the music industry. They might be producers, entrepreneurs that are sort of looking to push boundaries. And before we go really nerdy into this, because I want to go nerdy, I'll just have a more like consumer commentary if it's possible for you to do.

I briefly mentioned platforms like Suno before we got on. Could you for me dissect at least some underlying thoughts about the concepts of Suno, not specifically on the company, but about generative AI prompting making music and how that reflects into AI ethics, how we might feel about it and talk about it, and also the financial aspect. Wow.

Okay. Okay, let's begin. So let's, what we think about music, if we think about the music industry, so let's have a look at what we're dealing with first of all. So music industry, effectively the one that was created in the 20th century, and we still have a legacy of that now, consists of four things. Those four things are, you write a song, which is publishing, you record that song, the recording industry, you might play that song live, which...

performance area and you might sell a t -shirt associated with the artist who's performing live, which is merchandise. So recording, publishing, live performance and merchandise are the principal ways of which we all, regardless of how we reimagine them or we consider them, that they all pivot on music copyright. And so when it comes to AI, we have to consider what, how does that impact on those fundamentals.

So very simply, and we can do maybe a deeper dive into that in a second, in a moment. But when we consider, say, I'm only going to talk about generative AI today, just to keep it with such a framework. If we look at the idea of music creation, so if we think of that pathway from ideation to composition to writing of lyrics to, there's about eight processes through that.

There are a series of startups which are designed and all in competition to help and enable this term democratize that art form. Well, you can make up that what you like. And they are very exciting tools. I have to say, I love technology. But when we come to text to music generators like, you know, Suno or Udio and all the other iterations, because most of the large tech firms, it's generally understood, are sitting on the similar technologies. What does that do? Well, it cuts to the heart of everything that we have in relation to employment within music.

And we can analyze that through a series of lenses and consider, well, what does that mean? But for me at my age, having worked through as many jobs as I've done, the part that keeps me rooted is what's the future of work going to be like? That's the kind of pivot on it. So if you want to unpack audio, we can. How would you like me to do that? I think...

The AI ethics is extremely interesting and I think we'll reserve it for like 10, 15 minutes when we get really dirty. I want to understand the financial impact because this is the question that most people ask right away. What's gonna happen? What do you think? Well, let's have a look at what's already happened. So where were we prior to the introduction of generative AI systems?

We weren't in a great place anyway in terms of employment. You know, the majority of people, worked in music, if you were a professional musician, you probably had to have two, maybe three jobs. One of those jobs being a musician that didn't really pay for itself. Things were pretty bleak anyway. But, you know, there was a general understanding that this would need to change. For instance, the idea of finding more equitable remuneration from streaming.

I was always under the impression that we should just find new ways of creating income because the value of music is completely undervalued and that's at the heart of where the issue is. If we look at how important music is to people's lives compared to the amount of money that's generated by it as an industry, then there's an imbalance there that's relatively easy to address. I think Shane Shapiro made a point.

that I felt summed it up when he said the music industry is worth 40 billion, Starbucks is worth 120 billion, no disrespect to Starbucks, but in terms of people's lives, music has much more potential to grow economically. Yeah. We had Shane on the podcast where we talked about this in depth. It is such an interesting thing. Also, the one I struggle the most about being an entrepreneur in this space is that I know the importance of what I do from a...

both the consumer level, but also like the underlying value of what it brings to people and how little the industry is. So everyone who wants to listen to that podcast, go listen to Shane. Okay. So the - It was a space that AI entered into. So a very, very fragile environment. I was particularly inspired by the kind of international response to COVID where - many interventions occurred to support the arts, specifically within music, because it identified that there was a general concern, unquestioned concern that the arts were

an important part of society and that they needed to be supported. If we replace the idea of AI, generative AI, and the actors that may control that and imagine that is... a version of COVID, then we're looking at a similar type of intervention. What is it we care about? How can we protect it? And how do we make it grow? And that's the key root of my thoughts and motivations. That's a really interesting way of looking at it is we're looking at a potential crisis when it comes to

the financial viability of individuals in the industry and that generative AI. that draws parallels to COVID. Let me only study some more. Is it because you believe that people, A, people will lose jobs and B, people will make less money from the same jobs they did before? Well, I don't think they were making much money in the first place. So we take that as a start, okay? But there's always been the impulse to make music to work collectively, to...

to make things happen. And I think that in this transition that is occurring on a societal basis, music is a great way to have a look at the wide implications of this. This isn't something that just affects music or the creative arts. In fact, in some respects, it's, you know, one of the funny parts of all of this is the fact that secure jobs, or in terms of what you're gonna do if you go to college, you know, it's questionable whether, the idea of becoming a computer programmer is a smart thing to do. And that argument, which is always measured in education between STEM subjects and STEAM,

where the A is art, that's kind of like, oh well, we shouldn't forget about that. I think the arts as humanities can be quite positive about their future relative to other industries. But that's just a question of us all being in the same shit. And in that sense of community, there's a potential for, if we can see the connection that each of us shares, then we can find common solutions. That's my hope. So in other words, if I understand you correctly, because people all recognize how this affects all

aspects of life, the support towards the humanity is going to increase. the human aspect of creativity and jobs and output. Yeah. I mean, people are a number of philosophers and a number of large tech companies are already talking about the future of work as post work and what it will be like in this deep utopia that's been promised. Um, and while that all may sound very appealing in some kind of sci -fi spliffs, um, induced, um, Oh man, it all sounds so beautiful. There is a gap between where we are now and any of those

forms of utopia. And of course, any vision of utopia has an inverse vision of a dystopia. So I think that we need to contribute to shaping what this future looks like and dealing with some of the tough questions. And music gives us an opportunity to do that. And I do believe that it has a privileged place in that. And it's not just because I've spent my life in music.

I think that's widely understood. So for instance, I was at the conference in London two weeks ago called Get Paid run by the PRS supported by the UK IPO, which is the intellectual property office. And Director General Darren Tang, who's the head of WIPO, which is the World Intellectual Property Office, got up to speak. I haven't heard too many politicians speak, but he was brilliant.

He got up and he said, listen, music's really important. And everybody was kind of going, yeah. And he goes, I used to be a professional jazz player. And everybody went, what? And he said, yeah, no. So when I was in college, I was in Singapore, I was a professional jazz player. And I loved music. And he spoke about all the music that he'd grown up with. And he said, I went to my dad said to me, he said, listen, Darren, there's no future in music. There's no work in music. You should become a lawyer.

So he said he went to speak to the band and the band said, yeah, your dad's right. Now this is the director general of WIPO and he says, music's important. There's something mystical, magical about it. We understand, we can't describe what it is, but we know that it's important. And he said, AI is a threat and I'm worried about it. And I went, okay, this is...

This isn't us just shouting into the abyss. This is like generally understood. And when we see tools like Suno, or AlbiSeed, tools like Udio, we have to look at it from many, many different lenses. And one of those lenses is that they're terrifically exciting. So I have to point out, I love technology. I've always worked with technology. So I'm not a Luddite in as far as going, oh.

I love vinyl or the future, it should be like the past. I love technology, but this technology is different and we need to examine why that is and what we can do to help shape it. It's incredibly interesting. One of the things I talk a lot about with people and now I'm talking with an expert, so I'll throw my own thoughts out there and I'll have you evaluate whether or not it's correct or have a level of saying this to it. Is when people ask me about the future of music and artistry based on genitive AI, one of the things that I say is that I believe there's a movement towards the human connection much more than there's been the last five to 10 years.

When you just look at it from a macro level with the super fan economy, you know, sort of booming or at least projecting to boom. It's also a response to maybe a few years of detachment in general to music from a personal standpoint. And with, you know, social media, TikTok and all these kinds of things become overstimulated. We as human beings on a general level seem to be more and more attracted towards the human connection. So my assumption is, that as Joseph AI is progressing, we'll have an even larger movement towards the human connection.

That is, for example, having a close relationship with your fandom to an artist or being more aware of what kind of brand you're listening to, what kind of purpose. Is that a trend you're seeing or does it just mean making up sort of what's happening? No, there's lots of data on that.

Of course, with my brain, it's going to be anecdotal, so I'm not going to list you what papers and stuff, just things that I've seen that support what you say. So one of them is that, oh, it's Stanford. There's a paper from Stanford that showed that people prefer music that's made by humans.

I might be very reassuring, except there's a little caveat to that, which is that currently we prefer music that we're told that is made by humans. And I think fashions can change and it's quite possible to imagine and we shouldn't be scared of the idea that in the very near future we might prefer music that's created by non -humans.

And that might not just be a fashion thing, it might be something that I could be passionate about or you could be passionate about. So that's subject to change. I don't think we can see anything that's written in stone, but we can build upon the idea that we like people to work and we know it's hard to survive within music. And so there's a deeper dive below that into how are actually people being paid.

In terms of music and how do we grow the music industry into a music ecosystem, into a wider form of remuneration. And that's the part we have to kind of lean towards. I don't think we can rest assure that the idea of acoustic guitars and folk music is going to save the day. I, you know.

This sort of makes a quite natural transition into AI okay. As far as I understand AIOK, AIOK is a stamp of approval or concept of supporting the humanity and that music that is made with AI is trained on clear copyright material and respects the underlying human effort put into music. Is that correct? Yeah, it could be. So I'll just say exactly what it is. So when I was doing my PhD and I started looking at the future, It was pretty bleak. And there's a certain, I'm in France at the moment, so there's a famous quote,

I pray, moi l 'alluge, right, which is after me the flood. So when you get to a certain age, there's the idea that, well, the next generation, well, I'll survive, you know, when I'm dead, whatever. I was just thought, you can take a certain satisfaction from, looking into a bleak future as you're reaching your own mortality. And I went, is that the best I can do? And I kind of went, well, what's the solution to this? And as I was, at the time I was working with the IEEE and I came up with the idea that if there was a mark, a trust mark

That represented what stakeholders on the sector specific basis, in this case, the music ecosystem, defined as being ethically correct, which means it's beyond the law, it can include the law, but it can inform legislation, that a number of people would prefer to buy products and services that were marked as AI okay. They had that. It didn't mean, didn't, what does that mark mean?

Well, that's a different day's issue, and that's to be based on what the values are. And what are the established, I have many ideas myself, but I'm very careful not to impose what I think is the correct way, but really to give hope and a good voice and a way for people to become, and by people I mean everybody who works within music or participates in music, to help shape what it is that the future should look like and what we want it to look like. And so that very romantic idea, It is a very romantic idea. I bought the industry about three years ago and it

Was met with a resounding silence. That didn't surprise me, but I felt I had done my job. And so I started researching philosophical tracks and that was very enjoyable. And then I got a phone call about 18 months ago. I think it was about 18 months ago. No, coming up to two years ago.

From some people who said hey, do you remember that stuff you were talking about with AI? Okay, is that what's happening now? I know it's gone. Yep So from there it began and so now I'm funded by Enterprise Island We started six weeks ago. I have a team And we're working around the world Within the industry we haven't officially announced yet. There's a website that's up and To try to figure out how we can do this and there's a number of approaches, there's a technological aspect to it. But the hardest part is to get people together in a room and say, what do you want the future to look like?

How do we proceed and let's market and let's help shape the future. Nice and simple. I think it's so interesting to hear about that 10 year journey. And you know, it's only like last 15 % of the journey where this is actually starting to make sense to people. And you've been not only, you've been foresighted obviously, but it's just through deep study.

Seeing what is evidently happening and the industry not being aware, but also who could have predicted, maybe you could, but I also understand that the industry would have not understood this three years ago. Yeah. And it makes sense. You know? So I'm curious, what is AI not okay?

Well, that's to be decided by the stakeholders. So it's quite, what I mean by that is that, and that's, I would have personal thoughts about what that might look like. And we could kind of, if we step back into history a little bit, maybe we could consider things that were technologically probably not okay. So one of those for me, And this is an example of legislation having too heavy a hand. So back in 1996, one of the unforeseen consequences of the DMC Act, which was the Digital Millennium Copyright Act,

Was that was originally designed to protect the film industry. And from that idea of what was called this internet superhighway, I think Al Gore gave it that term. And the idea was that the piracy of films was a concern of Hollywood. So the DMCA was brought in to protect the use of digital rights. But one of the unforeseen consequences was this idea of safe haven, safe harbor, which meant that, for instance, if you were uploading material, then the platform was not responsible.

For the content. And this little loophole, companies like Google and Facebook have admitted, allowed them to have a competitive advantage and they were able to build their platforms through this loophole. And so we ended up in a situation which has now been addressed where the largest music streaming, music jukebox in the world, which was YouTube, was effectively a pirate radio station. And...

That's, there are many, it's a fantastic resource, the fact that we can listen to whatever we want to listen to and we can find it. And the industry has worked its way to find equitable ways for Amazon Music and Spotify and all the rest of that stuff. That took time. The speed of AI development doesn't suggest that we will have time to legislate.

These changes, these emerging behaviors as they appear. And so for me, what would not be okay is something that we didn't foresee and that we had no way of responding to. So on a very simple high level, but also actually this is pretty much it, AI okay is kind of like the bat signal. And so it doesn't matter whether it's the riddler or whether it's the joker or you know.

You just dig it up and you go, okay, this is not cool. And part of what the technology we're developing is that's a revocable label. And I won't be the one who's controlling its use. And we've done a deep dive into this. The people working with me on this have helped create DDEX, which is an industry metadata standards agency.

And they've tried many times to do different things. And so they have a vision of how a... body can be developed that will actually help create the standards that would define what is okay and what's not okay. But I'm excited about the fact that it can be in real time. So if someone comes up with something that we just go, wow.

That we can just not be kind of block it, but we can at least immediately identify to stakeholders. And by that, I mean people who care about music that something's not right here. And then that technology could potentially be fixed or just like an alert system. So that's the big dream.

I think that's incredibly exciting because as I research AI myself a lot, I'm a very practical guy, so I try new tools, two, three, four new tools every single week. So even though I don't research as deep as you do, I have a very practical understanding of what's developing. And I think it's scary. I think it's scary for the livelihood of the industry that I want to be in for the rest of my life. I'm not here...

By choice, I'm here by almost what I would define divine force to be in the music industry. Like this is my life, this is my identity. And I'm not gonna give up on trying to make sure that this is a good place to be in. And I think that's a very noble cause, but not even that, it's very practical cause that's gonna help us position ourselves for not making this a bad place to be in.

Tell me a bit what's happening with the support you're receiving or the interest you're receiving. You don't need to go into practicals if there's thing in motion, but like, okay, you were called up 18 months ago. You got funded six weeks ago. What is the industry reaction to the concept of AI okay or the broader philosophical thoughts behind it? I have to say, very positive. So I think I can say that I've been in conversation. So, The person who works with me closest is a guy called David Hughes. David Hughes was CTO at the RYAA in America for 16 years. He's based in Washington and he's the industry

Guy. So he introduced me to a number of people who at the senior level who work in AI and the recording companies and we're in conversation with them. There's clear concerns. From them, but the main part is that I'm finding it's very positive. So I'm hopeful that we will be able to make a series of positive announcements as we move through, because clearly those companies are in competition. And so this is not a cartel or any form of, it's just a way of making sure that whatever it is that we call a music industry,

Can transition into a future. So in America, a lot of work is being done and it's all going there. I'm also interested on the local level. So for instance, we have been in conversation with UK Music, which is the largest trade body in the UK since September. That's been very positive and that they are supporting us and we're going to see how that manifests.

As we move forward, but I'm very excited about what that will look like. I'm going on the road to Stockholm and then to Norway to Christian Sound where we met in May as well and then through Germany. And I'm going to see pretty much by the middle of June how many people are interested and how many groups in doing a meaningful approach to making this work.

If there are any philanthropic rock stars who are looking to create a legacy, this is one marvelous way that you can do it. Because I would love for this to be funded at a very senior intergovernmental level. And I'm in conversation at early stages with some of the people who I mentioned earlier in this conversation. And if we can make it work.

I really think that we can help shape the future. And we, not me and David and Sarah and my team, it's you, Jacob, and your listeners, and figuring out a lot of the tricky questions about what is it that we care about, how do we shape it? But the important part is the care is there. So I've spoken to a lot of small AI startups, and I haven't met one yet.

That resembles a Bond villain whose idea is, I want to wipe out. Everybody either seems to love music or is genuinely dealing with a technological solution that they're fascinated to solve and they want to make their business work. And wouldn't it be great to be able to communicate to that desire in small startups to say, hey, look, this is the kind of tools we want.

And people like Scott Conner are a good example of people who are finding new approaches to that. So I think that's all part of a larger communication system that's required. But the appetite is there. Absolutely. And there's something I've mentioned in the intro where you weren't on the camera, that's going to be at the start of the episode.

This is also a partnership of episode with Wise, which is a conference in Berlin, a think tank about me, myself, and AI people. If the audience comes there, they can meet you, they can meet me. I'm going to talk about a moderated panel investments in music tech and AI. And you're going to be one of the guests there. So we're going to have a couple of days talking about this in depth with different aspects to it. Maten, I'm going to ask you to be something that you probably be a bit.

Uncomfortable with, but I'm going to ask you anyways. If you were to be an oracle of a dystopian future where we did not address this head on, how would the future look? Really? Jakob Wredstrøm (39:37.934)Okay, so you know that Lenacow and so you want to go darker. Is that what you say? I want to, I want to understand the depth of where this can go. If we do not take responsibility.

Well, OK, so my motivations for AI OK were because of what I saw, the dark stuff. I don't think about this a lot, so this isn't in any kind of hierarchy of shit. It's just, here are the things that would inspire me. So when I first did my research, one of the things that I noticed, which was that companies like DeepMind and OpenAI, who were making the most innovative tools, and at that time, which I think was 2019, 2018, I just went to something different about what those tools are. And I was right, because those tools, then became the foundation models of large language models that we can now see how it

What I wrote about though was that I noticed that I was wondering if there was a connection between the business that they were in and the shape of their tools. So the business that they were in was this thing at that time, which was called AGI, which was, you know, best summed up at the time by Nick Bostrom's book, you know, Super Intelligence, and that was 2014. So that was pretty bleak.

This idea of AGI, because to me that had this sense of a certain religious honesty to it. Do you know that man will create God or whatever. That's become quite mainstream. And OpenAI are still in the business of AGI. And most of the forecasts for AGI are pretty scary. So...

Like Ray Card's father we know from music. predicted that the singularity, he's got a book coming out in June. I think he originally put it at 2040 something or other. It was revised to 2026. Other people think it's going to be 2025. These are people who I tend to think are correct. And so I go, oh, all right, AGI.

So, what does AGI mean? Well, AGI... If you do a chat GPT search for it, it's interesting because OpenAI redefined what the notion of AGI was back, I think it was last October. They didn't get much attention, but it caught my attention. And it was like, what is AGI? So for them, AGI was the ability of a medium coworker. So I asked what was the medium ability of a medium coworker.

Into Chach -EBT for electronic musician. You might try it as a jazz musician. And it gave me a breakdown. And it looked like a lot of the people who I know who work in music. And I was going, oh, this is kind of interesting. I'm worried about the idea, for instance, in Ireland, we have the world's first basic income support scheme for the arts. I...

Don't like the idea when we hear about the idea of universal basic income because I wonder who it is that is suggesting that. I do believe in democracies. I believe that competition is good and that we need to support the arts. But there's also competition and vitality and coming up with new ideas and getting people into a club and trying to get your record to be popular and...

You know what I mean? Like the lifeblood of that. So at the heart of all that is the sense of purposeness, like why do we do anything? So the darkness, the darkness comes when you're faced with why is it that you do what you do? And I think for a lot of musicians, the experience of using audio is kind of heartbreaking.

And I can only say that, you know, anecdotally, but I can also say that happened to me. I was using Ableton 12 and I was really enjoying it. I was doing it my evening times. And then one evening, I never mix alcohol with music, but I had a glass of wine and I went upstairs and I had to listen to, I want a glass of wine, and I went upstairs to listen to the track I was working on.

Which is gnaw, like anybody who works in music. You don't listen to the mix when you go home, you listen to it the next day, because it's going to sound like it's crap. We all know these things. And I just turned on audio, which had come out three days earlier, and I was astonished. And I went, OK, maybe that's just me on this gnaw, maybe it's time to hang up, which is probably true. But I spoke to some other friends who had similar experiences from a range of genres.

And I just went, this is dark. Like, does it become something that we give up on? And I know, so I would worry about that. And I think that's the fundamental kind of thing. Yeah. I had a similar experience a couple of weeks ago. So when I deep dive into things, I hyper -focus. So, okay. I was like, I'm going to figure out, uh, Suno.

So in a matter of weekend trying to combine the three kids and I was just sitting all night and doing it. So after 14 to 16 hours, I was just sitting straight trying to learn how to prompt the best way. I produced 10 songs from like full length songs, full structured lyrics that I knew could be on radio. The only differentiation right now is some pronunciation issues and then.

Audio quality issues that has like some frequency things, but like, you know, I'm an audio engineer by trade, I have a master's in audio engineering. I could quickly figure out at least two how to, you know, decrease that problem. So after 14 hours of work, obviously I come from music and I've done this for a long time, you know, but anyone who doesn't come can maybe do it in a long time. I create something that I knew like this could be on radio. And for me, that was just a very interesting thing. Like, You know, I have worked with producers that have, you know, been developing for five years and I know like right now I can prompt better than you can do after five years, generally speaking.

I don't know what that implies. It's just very, very interesting. What do you think it implies? I think it implies that we need to rethink the notion of what it means to create music. I think there's this pursuit, at least, you know, I'm a producer and engineer by trade to, over many, many years, spend time on becoming good at your skills. And then if you select, can use those skills to make great art. I think the process is going to change.

Where the quality of thought, the quality of intent, the quality of have something to say is going to be by far the most important thing. The tools of which you get there can be many fold and it does not necessarily need to be some you have built yourself through many years. So I think that is what's going to happen. I think that the thinkers, the people reflect that can tell stories that resonate.

and knows how to present it in that way is gonna be the ones who are gonna create music in the future. Yeah, I tend to agree with you on that. I think that's actually very positive. Because one of the things there is that if you take that description that you gave and you could apply that to pre -20th century music, like how do folk songs begin because they were good stories or they had the melody was the only work through.

And so I really do agree with that. And I do think that these tools can be really exciting and they can really help and shape and there's lots of positive aspects to it. It's just a question of transitioning between where we are now into what that future looks like. And that's the one thing that's very different about the new technologies is the speed of adaptation.

So outside of the affordability and the scale and everything else, it's just a speed. And so when I look at the history of music technology and we think about re -imagining a piece of technology, whether that's the saxophone or whether it's a record player or 808 drum machine, there's a time lag or culture lag between where some figures go, no, that's useless.

And somebody else picks it up and goes, oh, hang on a minute. What if I do this with it? That doesn't happen with these technologies. And as a result, it's kind of like being in the car with someone who's drunk. That might be an exciting experience. It might be the something that you hopefully live to tell your grandchildren they should never do. But it's not necessarily.

The wise approach and that's kind of where we are. And so that's where ethics comes in, which is to say, well, you know, hey, that may be legal in certain countries, maybe not speed limit or whatever else, but just to extend this metaphor. But the point is we need to figure out what is it that we want and how do we shape the kind of future that you just described there.

Matin, we're going to meet in Berlin. And quickly. Yeah. Sorry. And quickly. Yeah. And quickly. Quick. There isn't a whole lot of time for this to be. I agree. I agree. There's so many conversations to be had this. I have it on a daily basis with my wife also, who just, you know, what's happening to your industry. And we want to talk about this at Wise. And I'll see you there in around three and a half weeks time, I think it is. Matin, thank you so much for your time.

Lending your opinions and your experience into this. I believe personally that the topic that you're working with is going to be the most important the next years to come for the music industry. Technology is going to happen. You know, that's just market forces. Opportunities would present itself, but we need to proactively, consciously take an effort on the things you're talking about. That is not going to happen automatically.

It's going to happen by intent and effort. Yeah. And in summary, one of the key things here is that one of the reasons it's called AI-OK is because it's not about prohibition. It's not about banning or saying, oh, that's raw, or this isn't that. It's a bit like going in the case of Udio where they say, hey, we're just using fair use legally. You go, well, you might legally be able to do it.

Get away with that and legally you might be correct and a lot of lawyers may benefit from interpreting that for you, but no thank you, not here. Not today. We don't want that. Pay up or sign off. That's basically the mission statement on behalf of AIOK. So there you go. All right. Amazing. Thanks for having me.

I will talk to you in three weeks. Thank you so much for being here. We should probably have you back in a year's time and see what happened because this is going to happen a lot of things for you the next 12 months. I hope you're ready for the battle in the long run. Well, with help, with help. Absolutely. Talk to you soon Martin. Okay, bye now.

More from Sound Connections

All episodes