AI vs the Music Industry: What’s Actually Happening?

Jakob Wredstrøm

Normally I just wait 10-15 seconds and then... Is it summer in Ud? That's great.

Becky Brook

Re-ho! Yeah, it's no, it's only about 10am and it's already 27 28 degrees so well it's in my flat anyway.

Jakob Wredstrøm

No, really.

Jakob Wredstrøm

Okay, I think we're good with internet speed. So if you're ready, I'm ready. Great, okay. Hey guys, welcome back to a special episode of the Sound Conditions Podcast. One of my good friends are in the studio, Becky, welcome.

Becky Brook

Yes.

Becky Brook

Thanks so much for having me again.

Jakob Wredstrøm

Becky, we're gonna do a different type of episode and it's almost like expert commentary on movements in the music industry and I usually don't do these but I think it was so interesting to talk about this and there was no one better than you to talk about it with. So Becky, we gotta talk about some things happening in the music industry right now legally with music generation and major labels and I have on purpose and I disclaimed this before I went into the call, not research this in detail because I would love to understand it from a much more basic angle. So Becky, people can go listen to your episode from last time, but very basically just to set the scene, who are you and what do you do?

Becky Brook

I'm a music and technology advisor who focuses on predominantly on music rights. And I mostly work with startups who want to use music rights for their product or pitch services to music rights holders. But occasionally I work with large music and games companies if they have a sideline project that needs either music rights knowledge or tech implementation knowledge for music rights, that kind of thing.

Jakob Wredstrøm

And that's very relevant for today because we're to talk about... Well, maybe you can introduce what's happening and then I'll be listening and asking questions. So Becky, what is happening out there right now?

Becky Brook

Well, when we caught up late last year, we were talking about kind of, we talked about the high level licensing landscape. And then we also talked about sort of music licensing for AI and large language models at a very high level, at a kind of right at the beginning. And I think when we spoke, we said we'd catch up in three months. And I think we were hoping that maybe in three months, a bit more clarity. Well, we're kind of pretty much almost nine months since we spoke, maybe eight. And really we haven't got much more clarity. It's the unfortunate reality. Lots has happened and we can talk more about those movements, but realistically I think the picture's getting murkier in the short term rather than clearer.

Jakob Wredstrøm

And that is really the interesting part of this conversation today is I'm confused right now. So I've seen some headlines with Suno and major labels and potentially striking licensing deals. Could you explain to me what is sort of going on, at least on the headlines, the media, and what might that mean?

Becky Brook

So I guess there is the kind of major news in terms of music, AI, and lawsuits and or discussions. And then I think there is broader recent findings from AI, which we should also talk about more broadly, which are not music specific, which undoubtedly are likely to impact how we think about music and AI. So in terms of the music industry and AI licensing, The Suno and UDO lawsuits are still rolling on. It's going to be a while before we see the outcome of them. Same with Anthropic. And we should talk about the Anthropic example in more detail because that one, you know, there has been more movement on the music publishers suing Anthropic. And that's definitely like some of the things that have happened since we've spoken on that particular lawsuit are interesting to learn from. But still, frankly, we haven't really got any meaningful outcomes in the... in the music industry sense, so that is to say all of the high-profile large language models that are being sued by one or many parts of the music industry, mostly those are still open cases and so we're really having to look beyond music and look at earlier lawsuits from things like publishers, book publishers or authors. to look at early findings from these lawsuits. So it's a fascinating time. And I think in the meantime, what we're definitely starting to see is the suggestion that Sino and Yudhya are starting to try and license. Now, we don't know how serious their attempts are, but it's a really fascinating time because I definitely was quite taken, how do put this? It was something to the effect of... There was an announcement that Suno and UDO were starting to talk to the majors about licensing and it was like, and then there was definitely quite a lot of pushback from both the startups that have been trying to do it the right way from the start, trying to get licenses. And then there was also pushback, I think, from other camps concerned about what those deals might look like. And I think the former camp, the people who feel

Becky Brook

So somewhat cheated that they tried to do it the right way and possibly they're still not licensed. And then you have those who like Suno Nudio or maybe even Anthropik who didn't make any effort to license their rights and now are getting a seat at the same licensing tables and they to somewhat cheated. And I was kind of somewhat surprised. I was actually quite surprised by that because it's naive. to think that if Suna and Udo are coming to the table that the majors can't engage. Doesn't mean they'll agree a deal. Doesn't mean it'll agree a fair deal. But the reality was as soon as those large language model companies want to talk about music rights, the music, the majors have to talk to them. They would be stupid not to commercially, but also they can't be seen to refuse to license, not least because they're suing them for not getting licenses. So if they're then shown to...

Jakob Wredstrøm

Yeah.

Becky Brook

cross-draite it and refuse to license, that would definitely long-term work against them in court. So it's a really fascinating time. I think it's understandable if you're a startup who's been trying to do it the right way, that you would be frustrated that you've been trying to do it the right way. You've probably spent hundreds of thousands or millions of dollars on legal fees and advisory fees.

Jakob Wredstrøm

Hmm.

Becky Brook

and you've probably spent a year or two or maybe even three trying to license these rights and maybe you finished or maybe you're still in the throes of finalizing those agreements and it's understandable to feel hard done by. That is to say, you know, we talked about this before, right, like in our previous discussion when we talked more broadly about music licensing, I think I said something to the effect of like, you're not really given a special deal for doing the right thing from the start. If anything, probably the precedent suggests that your deal is probably better for not licensing upfront and waiting till you're sued. Because if you look at how UGC services operate and their commercial models for licensing, because of the fact that they were dragged into the licensing, kicking and screaming, and they had DMCA protections to support. to protect them essentially in early years. Essentially, UGC services have preferable terms essentially over normal streaming services like Spotify. so, it's understandable if you're a large language model or another AI generative model who has been making a concerted effort to license from the start. It's understandable to feel a bit hard done by frankly, because the music industry has never given you a better deal for asking first. And unfortunately, the precedent definitely suggests that if you can fight it out in court, then you might get a better deal on the far side. Now, I wouldn't say you always do, of course, there are plenty of startups who got sued for infringement and the settlement amounts bankrupted them, right? So it's not to suggest that you should always... I mean, it's not for a moment to suggest you shouldn't license upfront, but more to say that it's understandable if you've been making the effort to license and investing a ton of capital into getting the rights, it's understandable to feel a bit hard done by it by the fact that, you know, if Suno or UDO gets a license in six months time, they will have done it much quicker than the startups who tried upfront. And, or they possibly might get better terms because of the fact they've gone through the court process and they've robustly fought out what needs licensing and what doesn't.

Becky Brook

And so I think it was naive of those startups to think that these discussions wouldn't happen between the majors in Tuneo and UDO. I think that would be silly, but it's understandable to also feel pissed off about it. Like pissed off that you're not getting any progress despite being the ones who made the effort upfront to invest in the music rights.

Jakob Wredstrøm

Hmm and the sad part that really It justifies what you're saying is this is how a lot of the big companies in the market right now has done it and succeeded and is now sort of friendly with the music industry. Also, know, Spotify doing the way they did. there's plenty of examples where like you can't just, you know, play by the books right away always. And Napster and maybe I'm misinterpreting, but like I've seen this myself in my own studies. Like there needs to be this level of controversy to really get these things to change.

Becky Brook

If we look at the biggest music services available, I don't think it's true to suggest Spotify wasn't licensed. They have always been licensed properly compared to people like YouTube, who went down the avenue of hiding behind DMC, like behind user generated DMCA mere conduit protections, or musically or people like that. Services that started as user generated services were very lucky that they were able to least for a long time, hide behind legal protections designed for internet providers back at the turn of the century. And so I don't think you could draw, I don't think you could say Spotify was the same as that. Spotify did get licenses from the start, but you're absolutely right to observe that some of the biggest companies now, because TikTok is essentially, the base of TikTok is Musical.ly and Musical.ly was unlicensed too, right? So, you know, I think it's true to say that many of the biggest platforms now that we have in the music industry started out unlicensed. And if they hadn't started out unlicensed, it's very likely one or many of them might not have ever got this scale.

Jakob Wredstrøm

Let me just re-ask that question again, because as I said it, I was mixing it with a different thought. Is that okay? Obviously, I know that's the case with Spotify, so let me just try that again. And just basically to support what you're saying and also what I've seen over and over again is there is plenty of companies that's been disruptive and also done a big change where that has been the model that they need to ask, sorry, later and have somewhat succeeded that way.

Becky Brook

Yeah, go for it.

Jakob Wredstrøm

So what are the specific examples we've seen on platforms like that are big now that has gone down this route of asking for permission later?

Becky Brook

So I think there's a big swathe of companies that asked for permission later or ended up being sued. And lots of them don't survive those lawsuits, right? Because often the settlements are very substantial and often it completely clears out the capital they have and more. And it's very hard, to get additional capital if all you're doing is handing it over to the rights holders for historical settlements, not for growth. But if you look at the current largest platforms in the music industry. The reality is a good number of them started out without licenses, largely because they were able to operate under some degree of DMCA, mere conduit protections, they were argued to be user generated platforms where, just to recap on that, it's like the basic principle of those protections from the turn of the century was around internet service providers not being sued for their users. illegally downloading music. when YouTube particularly came around 20 years ago, God, that's horrifying, to when YouTube started, I think it was 2004, 2005, they were able to tap into those same protections because they were able to say, well, we don't know what our users are doing and we're not encouraging them to infringe. And for many, many years, that enabled them to scale aggressively because they didn't have to worry about music rights. be very clear and say YouTube and all of the other historically infringing services is very much a historical thing. But they were able to get out of the gate and get a good amount of scale before they had to worry about music rights. And that would be true of Musically as well, which then became what we now know as TikTok because ByteDance acquired it. so some of the biggest services built scale before getting music licensing. And you can see why that's appealing.

Jakob Wredstrøm

And again, you know, now we're sort of guessing, but I think, know, when, when Suno, for example, got this rather large, funny round last year, I believe it was, I'm sure that's what some of the conversations that's been going on in those rooms is just like, okay, we see there's a president for asking for permission later and Now we're going to try to create the best product ever and get some leverage of public sentiment and maybe play in courts until we get to a point where they need to settle. I can only imagine that would be the most natural assumption to have.

Becky Brook

For sure. think it makes a lot of sense to string this out at the stage they are. Strategically, it makes sense for them to start dipping their toe into the licensing talks. But the reality is they're not going to enter into substantial and widespread licenses until there's a bit more clarity from the current suits, right? And I say that, that's not to suggest they might not settle before their own suit is resolved, but until there's real clarity and probably in the case of the question about do you need a license to train a large language model on copyrighted materials, realistically that's probably going end up going to the US Supreme Court. And so, you know, it makes sense for them to seem to be acting in good faith by talking to the majors and similarly the majors need to engage in those discussions. But why at this stage would Suno hand over a large portion of their cash and their equity unless they felt they had to? I'm not suggesting that's moral argument, by the way. It's definitely not the moral argument. I guess the unfortunate reality is, I'll say that unfortunate reality is probably a little bit aggressive, maybe we'll scrap that little bit. I was going say something to the effect of, you're a Scandinavian, I'm a Brit, we both come from more socialist leaning societies. But the reality is these large language models are based in the US. And let's be clear, all of us, both the UK and pretty much all of Europe and definitely the US, these are capitalist societies where

Jakob Wredstrøm

Yeah.

Becky Brook

If you're a incorporation, your obligations are primarily to your shareholders, right? You know, and so the reality is why would the shareholders don't want to put the money into this either? They don't want to put the money, know, investors don't want to generally put money into buying rights. Like that's why there was, I think, who was it? Like, I mean, Sherry, who does a really, and Water Music do some great data around investment in music tech and But there was a separate statistic, and I can't remember the source, unfortunately, but I should probably try and dig it out, where essentially someone was able to show that licensed large language models raised something like 30 % less than unlicensed large language models. Because it shows that, and let's be clear, those ones that are licensing copyright are generally raising smaller amounts too. Like generally, I mean, it's in there, in setting out to raise smaller amounts. Now look at Suno, Udo, OpenAI, Anthropic, those are all unlicensed, right? And they are all being sued by one or more copyright holders. And they are the ones that have raised all the money and they're also backed by the biggest tech companies. So, you know, it's understandable if you're working towards the interests of your shareholders to not want to license music. But the moral argument would say, We can't have ultimately large language models or machines learn off human creation, create works that can compete with those human created works and then savage the market for commercial music and therefore the value for human creation. so, you know, morally, I don't think it's right, but you know, the reality is these companies are incorporated in the US. And it would be very hard for them to do something different, frankly. But then what we have to look to is the law. We have to look to our legislators to sort of make sure the laws are robust enough to protect against this. And that's where things have really become very interesting in the last few months.

Jakob Wredstrøm

Hmm. I have a question that I've been thinking about lately and it's sort of trying to separate the emotional reaction to the logical reaction to what's going on. Because what I experienced, you know, on my LinkedIn environment, which suddenly has become a very toxic LinkedIn, I don't know what's happening to your algorithm, but... is when people are reacting to things like, for example, Timbaland going out and launching a new AI label and he's been supporting and advertising for Suno for a while. I see lot of emotional reaction as by far the strongest aspect of it. And I see less and less so factual, logical, mathematical statements about what's going on and how that's going to damage the current industry. And obviously there's going to be someone there fighting that battle. But I also perceive A lot of this being like either conscious or subconscious reaction to the threat of quality that Tsuno is providing, like the livelihood of music and that sort of overshadowing the emotional landscape of the rightful allocation of payment to some degree. Like I don't really have a question associated to it, but that's sort of the angle that I'm seeing in real time, if that makes sense.

Becky Brook

Yeah, it does. mean, I can see why many of the arguments are emotional, because particularly if you're a creator or you've built your livelihood on making music or supporting human creators, I can see why there would be an emotional response. You would understandably feel cheated, because to your point about equality, right? Let's be frank. these outputs would suck where the input's not good. I mentioned it on the last podcast and I can't actually remember whether it was Suna or Yudio who said it, but one of them, when they first launched, obviously there was an immediate backlash and one of the CEOs said something to the effect of, we're not saying we licensed on commercial music, but if you want the best output, you have to put in the best inputs. And it's true to say, it's like, large language models, they've leapt ahead in terms of how good the outputs are. They've got infinitely better in the last 10 years as we've discussed previously. But if the inputs were just kind of production music and or public domain music, then the outputs would be far, far more limited in their quality and their scope, right? know, public domain music is largely 70 something. well, 70 to 100 years old, right? So there's not going to be anything in modern styles. And if you look at what's production music, maybe excluding people like Song Trader and Epidemic, a lot of it's functional music, it's lift music, it's mood music. And so the reality is, I can understand if you're someone who's built your career around making music, I can understand why you feel cheated by these big companies, but I do also agree with your point about, I do absolutely agree with your point about the fact that ultimately there is a strong logical pushback as well. don't have to just do, it's not, you don't need to rely purely on the moral and emotional argument. You can simply say, well, this is copyright infringement. And I should say, I would want to give Ed, Newton, Rett's credit for that.

Becky Brook

He has done a good job in the UK in particular fighting, not just the moral argument, but also highlighting quite how clearly, in some cases, the of Cino and Udo, the work he did for MBW, quite how clearly the inputs are derived. Can you hear the background noise? Sorry, I realize I can definitely hear it. You can't? OK, I will start that last sentence again. I was going say something like... It's definitely worth giving Ed credit for the work he did for MBW, analyzing the Suno and Udo outputs, because he really clearly showed that the outputs were derived from the inputs if he used the right prompts. so, you know, I think it is, I think the industry has for too long taken a rather moral high road rather than a sort of analytical

Jakob Wredstrøm

Hmm.

Becky Brook

and factual approach to this. I think, I don't think I said it, I'm not sure I said it on the last podcast, but you know, the fact that the industry has for the last five years or so, sort of spent a lot of time talking about the superiority of human creation. And that's why copyright only benefits human creation, not machine creation. And I think that's a naive, I mean, I've used the word naive quite a lot today. But I think that's a wrong-headed approach. think ultimately, if, and it's a big if now, I think it's increasingly, the if is becoming bolder over time. But if it comes down into major jurisdictions that you need a license to train, why would you put your big commercial music into a large language model where the outputs are free of any copyright? Because as soon as they're created by the end user, made available in any form, then they're public for everyone to use for free because there's no copyright on them. And so I think it's incorrect to take this kind of approach that human creation is so much better than machine creation. morally, it's the only thing that can benefit from any kind of copyright because if and when they are going to license Suno, Nudio, et cetera, what's the commercial model if the output's are free of copyright. know, the commercial model, of course, we know will be equity and cash upfront and some payments based on what users are paying to create the music. But if there's no way to commercialize the outputs in a meaningful sense, then the opportunity will be far less than the risk for these rights holders. Does that make sense?

Jakob Wredstrøm

Hmm.

Jakob Wredstrøm

It does. It does make sense. And that's also the confusing part when these investors came into. Sunu, for example, is just like, I was very complex too. For what end? Like, what is the imagined end where the investor is going to have return on investment? And that's for me, like the big question, because like, there's so many un-clears in this. what's going to be besides, you know, SaaS at scale, like... What is the model there? And obviously, when you go into this battle that there must be more in order to make an outsized return investment on $250 million investment. What's your thoughts about this?

Becky Brook

So I guess specifically on the So part, they have said publicly that, you know, they're confident they have enough money to pay the settlements if they become due. And I think that's probably largely right. I mean, I think they underestimate how big settlements could be, not least because if it's, you know, they end with statutory damages in the US, it would be a wild amount of money. But they're not crazy. If you look at the history of things like YouTube or Musically, that the suggestion is that actually if they raise enough money, they will be able to scale out of it and they might get better terms on the far side. And so I think when you think about how the investors invested in Suno and Udio, I can see why you would invest in Suno and Udio because logically, if you're a capitalist and your intention is to make money, the there is a meaningful chance that one or both of those companies will do quite well. And I think it's already been proven that end users are now creators, right? Like I think I state, I use the media statistic of just 30 million people who make music in some form, let's call it some semi commercially, they release it, whether through a DIY creator or through a major label. As 30 million people who create music. But what apps like Instagram have taught us is Everyone is photographer and what we now seem to see is everyone is a video creator, as TikTok has shown us. And so, you know, we don't yet have a world where everyone is a music creator because let's be honest, until now, you needed some skill to create music. That's not to suggest that all of those 30 million existing music makers had much skill, but they had enough technical skill to make something and release it. And when you look at people like Suno and Udio, they are moving to the world where you won't need any skill to make music. And so this kind of idea of the Instagram model, that kind of removing all barriers to humans creating quality content is...

Jakob Wredstrøm

Hmm.

Becky Brook

is well understood by Silicon Valley, by VCs. They understand the idea of everyone wanting to be a creator. therefore, an AI is how they can see music becoming of barrierless in terms of creation. It's already a very low barrier, to be clear. Like we've moved as an industry. You when I first started, nearly 20 years ago, you needed skill and probably access to capital to release music and connections as well, which is why major labels had such strong market share because you know, stores had limited shelf space and so, you know, and studios cost a lot of money and home studios weren't yet a thing really in a meaningful sense. And so you think about how hard it was to make music 20 years ago and release it and get it to the end consumer. And you think about where we are now with 30 million music. creators and you think about like now you only need skill. You don't really need any money, right? These DIY solutions, these home studios, they are so good and they're so cost efficient. You know, you can do it. You can do everything with pretty much nothing. In fact, if you're willing to share in your income, you might need to put nothing upfront now pretty much. As long you have a little bit of skill, I say a little bit, not sorry, in musically brilliant, but just, you know, you have some skill, but you know, 30 million. compared to the idea of having hundreds of millions of people creating music or even billions of people creating music. Because then you look at things like Instagram or like TikTok, how many people are creating content on those platforms. It's not a crazy thesis, you know, it's a well trodden model for the venture capitalists.

Jakob Wredstrøm

But again, if you go back to the Instagram picture and you say, well, everyone can be a photographer, but to come back to brand and followers, like not everyone can be an influencer. And I think that's sort of where, where Asta Street comes in and, you know, Hope for Music is, yeah, we might be flooded by, by quality, like even high quality, but, but at the end of the day, what's always stuck is the icons, is the big storytellers and, you know, I've been speaking with a lot of people lately, especially here on the podcast, think episodes to come as well, about branding, about storytelling and what makes stand out and what is the... What is the grounding of belief of artistry still being able to go on when these things are happening? And it just all comes back to the storyteller, the people able to follow. And my skepticism to the drama around the impact of music, not from a financial perspective, on a creative perspective is also that, okay, great. Well, not great, but things have changed and that's disruptive, sure. But the base premise of being an artist is still very much there. And that is the connection over time, consistent narrative, having a personal history. And for example, with Suno, if you just made a comparison to Instagram, if a person on Instagram posted one type of photos that way, that day, and then as... completely different another day and then something different a third day, you won't gain a lot of followers over time because you need to have a consistent narrative. Like that's the whole thing with influencers, even if it's just like professional influencers or just consumer influencers. And that's the same thing with artistry. Like, you really believe that, and you haven't said this, but I'm asking the question, do you believe that artistry is in danger from what's happening with music creation? Or is it entirely something different that is in danger?

Becky Brook

I think as the barriers to music creation have reduced, it has flooded the market with music. And it is probably true to some extent that it has made it even more challenging for those professional creators, music creators we love, to make a profit off their music. But because I guess, know, said before, our time is, as humans, each one of our times, our days are finite, it's 24 hours in every day. And we have access to so much more music now that even the people we love probably get ultimately less money from us. And that's a perfect storm of lots of things, right? Like that's not to suggest, it's not purely because Spotify has production music and therefore people now listen to more mood music. Like we think we discussed the Stim writer who was number three top writer in 2023, who was writing for Epidemic Sound. It's not to suggest that's sinister or any, it's not to suggest users are incorrect. Users listen to more music now than they ever did before. But on a per capita basis, the money we pay for music hasn't kept up. look at people like myself or maybe you, don't know. When I was

Jakob Wredstrøm

Mm.

Becky Brook

Buying music 20 years ago, I was buying maybe 50 pounds worth of CDs a month, so didn't have a vinyl player then. And then I was also paying for OD2, crazy, for my time-limited downloads. So I was probably, as a poor student, was probably spending 65 a month on music before I considered live music. That was just for my recorded sort of products or my distributed music. And you think now, you know, it's hard to spend... as much money. Like what I do now is I try and patronise my favourite artists aggressively. Thought, this is quite a funny example, like Little Sims over here, that's her previous album, but the other wall has the latest album. I even bought a CD, I don't own a CD player, but I felt like my natural spend now, I've got Spotify serving all of my needs for music, let's be frank, it is Spotify in my case.

Jakob Wredstrøm

Thank

Becky Brook

And I only have to pay them, well, I think I'm paying them give or take 12 pounds a month because it's just me in my home and, you know, I'm not going to cheat the system and have a family account and give it to my friends. And so, you know, my 12 pounds a month spend feels insignificant. And I don't feel like I spend enough money on music ultimately. So what I then do is go out of my way to buy pre-order vinyl and buy the special edition ones. And in this case, the Little Sims Lasers release, I bought the CD because I was like, let me buy as much as I can because I wanted to. Sort of increase my spend on music. But the reality is like, you have to make a really concerted effort now to spend what you did equivalent to what you did before 20 years ago as a music lover. Even the end users, you look at like the mass market who never were big fans of music, who maybe listened to the radio and bought a couple of now CDs and a couple of compilations and maybe one artist album a year. That's probably still more than they are paying for music now. And if you look at it, if you look at it in terms of the cost of cost of capital, the combination of inflation, et cetera. They were probably spending on average maybe 10 pounds a month on music, maybe five on average. And if you look at what that is now, that's the equivalent of them spending probably more like eight. Well, we're not getting eight pounds out of the average music consumer anymore, right? So I think we, this is not to suggest it's anyone, any specific person's fault, but is it harder for commercially successful artists that we love to survive now than before, yes. And it's not the fault of any one thing. It's a perfect storm of factors, globalisation, digitisation, you know. And so all of that's true. But then when you add in the risk of AI, I can see why people are seeing this as a black hole, frankly, because, you know, they're already struggling to make money with the existing models. And live a sort of sustainable life as professional music creators. And they're already struggling to compete against a sea of DIY creative music. And to add to the idea of like, well, this is how it is now with 30 million music creators. Well, what if we have 500 million music creators, but we only still have the same current streaming services, et cetera. Like we're gonna have to solve the revenue sources. We're gonna need to... Becky Brook (36:30.658)...Jakob’s LinkedIn of 50x our income as an industry if we're going to sort of dramatically increase the number of available musical assets, but not increase the income, not very, very substantially increase the income. So I can see why people are concerned. You I think you can see a world where all kind of UGC implementations use non-human creative music. And all background music is non-human created. And that accounts for a crap ton of the music we listen to as a society. And so I get why people are, it's not just a moral concern, it's also an economic concern. People who invested their life into creating music. And these are people we are fans of, right? Like bands we love are concerned about this, or artists or writers we love are concerned about this stuff. And I can see why.

Jakob Wredstrøm

Mm. I think that's a brilliant way of describing it and sort of nuancing it. And I think one of the big takeaways I have for these type of discussions, and especially while you were talking, I was thinking about is the perceived value and the actual value of music are both very, very low right now. And as the creation or the barrier of entry lowers and much more people are creating music, I think both of them could also be negatively affected. Perceived value of what does it to pay for music and the actual value of music because that's going to be a subsequent reaction to the perceived value of music as well from consumer behavior. And that delusion is difficult and obviously there's nothing we can do to just you know stop technology and there is a hope for the legal systems to make some sort of solution to this but in reality that's going to have limited effect. So you might disagree, but this is how I see it.

Jakob Wredstrøm

Are we fucked?

Becky Brook

We fucked, goodness. So, if... I'm trying think how to best describe this. Let's let me formulate for a second. So I was going to say something to the effect of when we spoke last time round. In major jurisdictions like the UK, US and most European jurisdictions, it felt fair to say that there was little clear existing legal evidence that would suggest that you could train a model without a license, trainer model and commercial music without a license, just generally copyrighted items. And so, you know, our existing laws in the UK and the US and most European markets to my knowledge don't provide strong evidence that you could do that without a license, right? And in some cases, like in the UK, they clarified that things like the, so that there was a, an exception for data mining. In the UK, our government clarified that exception didn't apply for this use case. So nine months ago, you would have said, or I would have said at least, it felt like there was a good chance that in major jurisdictions, it would be deemed you needed a license to train after going through the courts. But for the first time in my life, feels like copyright has become something that is hyper political. Now, lots of people say, well, it was already political, but actually that's not true. If you think about it really in big markets like Europe, UK, US, depending on how you include Europe geographically, UK and Europe geographically.

Becky Brook

Largely copyright was kind of a bipartisan issue with some anomalies like US radio where they don't pay royalties. Like major political parties in most of these kind of Western markets supported copyright in a decent form. But AI might be the point at which we diverge politically. And you look at the early few months of Trump's time in office and you look at his.

Jakob Wredstrøm

We can't Google flags any political names. If you just do that. Yeah, so don't use any names.

Becky Brook

Interesting. That's fascinating. Don't name him, but I can talk opaquely about the direction. Fascinating. Okay. Good to know. Okay. So let me start that again. To the... I think I said something to the effect. I'll just recap where I was. So I was saying something about copyright being largely a bipartisan issue historically. Well...

Jakob Wredstrøm

Yes. I don't use names. Yes, yeah.

Jakob Wredstrøm

Yeah.

Becky Brook

Times are changing. I think in the US, the last six months have shown us that maybe copyright protection is no longer a bipartisan issue. The sacking of various copyright office executives, the public statements of powerful political people, very generic description, have suggested that some of the people in power in the US particularly, but probably also in Europe, are not fans of copyright. I think there was one quote which was something to the effect of like, fuck to your point, fuck IP. Irony of that person is they built their wealth on IP, just a different kind of IP. They built their wealth on trademarks and patents and what they actually meant to start copyright. But that's because they wanted to use copyright free of licensing. I think to your point earlier, there is a... I disagree to your earlier point about our lawmakers, our courts going to save us. We have to compel them to. Have to use a moral argument and a commercial argument to compel them to. But nine months ago, it felt like it was probably going to come down in the industry's favor. Was going to. Probably. The law largely suggested that they needed a license. The reality, unfortunately, is that you look at the report that the Copyright Office rushed out that said there was basically a need. Basically, you have to get a license. And then the Copyright Office head was sacked. I think the reality is it doesn't seem like there's much appetite in the US. To require licenses. So, you know, recent lawsuits that have been found in the last few weeks have mixed signals. But I think we need, we as an industry need to make a stronger moral and commercial argument to our respective political leaders about the importance of protecting human copyright.

Becky Brook

And I realize that could sound like it's contradicting what I said earlier. I'm not suggesting that, again, machine output shouldn't have some kind of protections on them. But we should be looking after human created copyrights. Otherwise, I think it is we are fucked as an industry. Yes, I do think that's true, unfortunately. I don't think the game is over. The law is not yet clear. We haven't entirely. Decided that is properly, there's no Supreme Court ruling. And even if it was determined by the Supreme Court, you don't need a license to train, which is possible. If the Supreme Court in the US decides that the use of inputs to train models to create outputs is a transformative enough use case that they don't need licensing, like what was described, the Google Books example. So if it's considered transformative enough, they don't need licenses. I think that's true, if they did, well, okay, so that might be true based on the current law, but then frankly, the music industry would be crazy not to then be lobbying to change the laws, right? So think this is going to be a really, really long standing battle, but... it is understandable to feel very concerned about the future. And I think we've got probably a couple of years to turn this round or otherwise neither of us will be working in music. And frankly, very few people will be working in music if we're not careful, other than those who manage the large language models or run the DSPs, which will probably increasingly be using music generated by large language models.

Jakob Wredstrøm

It's so refreshing hearing that take because that's what I'm thinking all the time, but it's so hard to... Let me say it this way. I... This should be misunderstood, but I'm on my way out of music when it comes to me thinking long-term industry-wise. I'm still very much present from a technology perspective. I still believe that there's much to be had and much to be won from that perspective. But I am so concerned and so confused about this that I'm out of music itself. And that's a decision I've made, and that's for sure. I'm not even half out. I'm out. And that for me is rather perplexing to say. But I can just see it like the insecurity of the future combined with the incredible insecurity of the present and also the past just doesn't add up, for example, for me as an entrepreneur. Like, I do want to say that I'm still here for technology. I'm still here for the surrounding elements of it, but... But the concept of music itself, when it comes to representing artists, songwriters, publishing companies, I'm not saying they're doomed, I'm just saying me personally, I'm out. Because I was so conflicted about the attractiveness of doing the work and the conversation that space that already now, like, I can't combine it with the insecurity of the future either. And I'm not trying to change anyone's mind on this topic, but like, I'm already being affected. And these are sort of the first ripple effects I also see in other people that's... Heck, is this where we want to take a chance? I believe that's the saddest part. I've called it out many times on this podcast that I believe that the capacity and capability of industry people could be significantly higher in general. And one of the reasons for why that maybe is not so is that it's not an attractive industry or understandably so at scale, at large. And these kind of elements does not positively contribute to that fact.

Jakob Wredstrøm

And that for me is the fuck part, if that makes sense.

Becky Brook

Yeah, I guess I don't think... I guess my assessment is there is, it is possible that we are fucked, but I guess you already feel like it is. I think we still have some time to turn it around. Think that takes, I'm still, I still think there's some light, there's still possible light at the end of the tunnel.

Jakob Wredstrøm

Yeah, I do.

Becky Brook

Because I do think the industry is passionate and smart and no one really works in music for the money. You work in music for the love of it, right? And so like, we would all earn more money working at banks or other things like that. And so I'm, I still see some light at the end of the tunnel, some possible light at the end of the tunnel. But, but depending on what you're doing in music, it like, you know, for example, some kinds of production music, I could see why people would be shutting down the hatches. And or if you are someone who has never managed to get a foothold, it's only getting tougher. So whether it's someone who has commercial success in an area that's already being savaged by AI or in an area that or in a career that never has really yet, it never has flourished, then I could understand why it feels unappealing. I'm still super passionate about it and I still think we have time to turn it around. But I do think there's a lot of, you know, there's a lot of concern, whether it be, you know, our lawmakers changing the laws or sort of forcing the courts to rule that there isn't a license needed to train or whether, you know, in the world where we do license but the outputs are free of copyright and therefore we end up with sort of a billion tracks available on every streaming service. Like that is a pretty bleak world. Someone made a really compelling comment to me at Music Ally next last week, something to the effect of, you know, essentially what we probably need to do is have DSPs like Spotify, like Apple, etc. only carry human creation. As much as these are suggested otherwise, it's not so easy to be sure where the line is drawn because as much as they say otherwise, those algorithms are imperfect still. You can detect certain models, you can detect artifacts of AI large language models, but the reality is

Becky Brook

Do we want to have the current volume of tracks only 10x or 50x? I don't think that's good for anyone. It doesn't mean that music shouldn't be made, it doesn't mean it shouldn't be available somewhere. Someone was making a play for the suggestion that we should have two tiers of services, commercial music services and everything else. I'm not suggesting that is the solution. But I can see why people see a bit of a bleak future. I still think we have time to turn it around. But I do think we have to change our thinking around the outputs.

Jakob Wredstrøm

Hmm. And the most interesting part is, I get a lot of clever people on this podcast and some of the, people that I considered the smartest are some of our common friends. And a few episodes ago, we had two, two common friends on the podcast and two different episodes. The first one was Dan Fowler and the second one was Daniel Noagar, the professor. And it was very, it's a very interesting contrast because basically Dan and Daniel, are probably top five, the two smartest people I know at the music industry. And they were so on the opposite end of the spectrum of like when it comes to the outlook of the industry where Dan was very bleak. He's obviously incredibly intelligent and sometimes a bit hard to sort of comprehend the complexity of what he thinks, but very bleak about it. And then I was talking with Daniel, which is probably the most knowledgeable music industry guy I know. And I know a lot of people. And he was so optimistic. I think that's also maybe like for me, what I just need to navigate is like the people that I respect the most in the world also have opposite opinions or opposite outlooks. And that really signifies how unclear it is what's going on right now. And hence also our conversation today. Becky. I don't want to put you on the spot, if we want to put up some sort of visual spectrum of outlook of industry and putting Dan on one side that you know and Daniel on the other side that you know, what would you place yourself on that hope scale?

Becky Brook

I'm roughly in the middle. I'll oscillate in either direction, but I average somewhere in the middle because I share a number of Dan Fowler's reservations, but I also still have some hope like Daniel Norgard. So like you said, I know both of them and I often sit in the rooms with both of them at the same time. And I don't think I'm all positive or negative. I think actually I'm sort of about some things. I'm negative about some things, I'm positive. I think we still, like I said, we still have time to turn things around, but there is lots of challenges, lots and lots of challenges ahead. And so I would say I'm somewhere in the middle, but that's not to suggest I always sit in the middle. I'm not a fence sitter, as they would say in the UK, but I average out somewhere in the middle.

Jakob Wredstrøm

Hmm. And maybe my last comment on that is when I talk with Daniel, I sit emotionally with Dan Fowler, like reaction-wise. And he's much more intelligent than me, so he actually has arguments. I just sit there emotionally. But then when I listen to Daniel speak, and I know enough to know what I don't know, and this guy, knows so incredibly much, and I'm just like... Thank you for blessing me with knowledge and I'm inspired by the way that he argued for hope. But I can't seem to get out of this hopelessness, if that makes sense. And that's just the honesty I want to have through the podcast. If I want to place my faith somewhere, I place it with the hope. Intellectually. But I can't be there emotionally, if that makes sense.

Becky Brook

Yeah, that's interesting. Guess I'm very lucky that day to day I talk to startup founders who I guess you have to be a little delusional and sort of mad scientist to want to found a startup in music tech. I mean, I'm using that very loosely what I mean in senses like I talk to people who are wildly optimistic every day and that kind of gives brings me background. It tempers my

Jakob Wredstrøm

You

Becky Brook

Concerns, it tempers my reservations. And so maybe that's good. Maybe that helps give me that balance. I don't know, I don't know if that helps you at all. Know, talking to startup founders who can't see all the barriers, can't see all the hurdles, can't see all the problems. As long as they're willing to understand those things, that working with them kind of

Jakob Wredstrøm

It does.

Becky Brook

Gives me quite a lot of balance. You know, I'm excited because lots of these people are working really hard to deliver something truly innovative and exciting, whether it be an AI algorithm for multi-track mixing in the case of someone like a RoEx historically, or I've got a client now who building a Belgian specific streaming service and are really trying to support particularly independent created, probably now mostly major owned, Belgian. Artists specifically. You know, I find myself really passionate day to day about the products my clients are building. And then I spend a lot of my time, lot of my, I this before, a lot of my emotional energy trying to advocate and trying to agitate even for the industry to make just basic licensing easier. Because if we can't make that easier, if we can't drive a diversity of DSPs, for example, I think we're going to be in trouble because we double back to the point that these large language models that we're suing, et cetera, are largely funded by the same companies that also own the current biggest DSPs. And so it would be naive not to see some... some rephrase that, it would be naive not to see some strategy in investing in large language models that might provide you limitless free music. Might, that is the word for sure. And so if the industry can't encourage new DSPs, it can't encourage new entrants, new innovation. And I say innovation, I mean, this sort of separate from the generative type. Mean, like new services, new user experiences, whatever. If we can't encourage those things along, new, also pure play services, because Spotify still is largely a pure play service, increasingly they're trying to move away from being a pure play music service because it's so hard to make a margin of music. We need to foster a healthy underlying music industry that cares about human creation. And so I often feel a bit disillusioned, like you seem to at the moment.

Becky Brook

But I do also know that when I talk to the people who do these deals with power, most of them understand it. So maybe we just need to be sort of shouting louder and maybe to eat it to the most senior executives in the industry about this. Because in a world, I presented about this at South by Southwestern Austin this year. It was like in a world where we only have sort of five big platforms and they also have like major investments in large language models. And you don't need a license to train, I think it's a very, very bleak future. So we need to be doing everything we can to address politically and legally the AI models. But we also need to be building a more healthy, sustainable flow of innovation for services, for user experiences, so that we have more people innovating the space who actually care about music, who are pure play services. And thankfully, I'm lucky to work with people like that every day. And so it keeps me more optimistic, but I wouldn't pretend that things are perfect at all.

Jakob Wredstrøm

I think this is a good place to leave part two of our conversation and return to part three. Let's not say three months, let's say six months, let's see the timeline.

Becky Brook

Probably another nine, because if it goes to the Supreme Court, it's going to take even longer. Something else I was going to say and I can't remember what it was. Is it helpful to briefly talk about the last week of lawsuits? Because there's been quite a lot of precedent that might be... I know if it's like... I hinted at it earlier, but I'm mindful that in a week, recording this in a week where there's been two really high profile lawsuits... that are not music but tell us something about the way things are going might be helpful.

Jakob Wredstrøm

Think this will be out the 15th of July, so if you believe that still would be timely by then, then I think we can do it.

Becky Brook

Yeah, okay, give me one second. Now my door's just wrong. I'm going to answer it one sec, as we're not recording right now. Hello. Thanks, Jenny. My very lovely postie. Okay, so what I was going to say was something like, I'm just going to triple check something because I don't want to get the reference wrong and it's important to get these things right. I did have them in my mind, but an hour after talking to you, I've like... So...

Jakob Wredstrøm

Of

Becky Brook

Second. An annoying thing by the way is I had this phrase that I wanted to use in the call but I can't for the life of me remember I've used it a million times before so I had to kind of try and hint at this phrase and it was like driving me crazy. Was I saying? So in this case, sorry, now I'm trying to do, now my brain's like all over the place. So essentially there's a meta lawsuit right and there's also an anthropic lawsuit. Both of them, they're not music lawsuits. But they're probably worth mentioning because how everyone's, they're seeing them as like in like, they're seeing them as indicating where we're going. I don't think that's necessarily true. But essentially to try to go into it and just explain what's best.

Jakob Wredstrøm

So what I can do is I can say.

Jakob Wredstrøm

See the most natural transition here.

Becky Brook

You could just do something like, it's been an interesting few weeks, hasn't it, Becky? Something like that.

Jakob Wredstrøm

Yeah, but yeah, think thank you so much for being on Becky. It's been a few interesting works and maybe just in the closing remark before this call, we just discussed some things that just happened recently. Maybe you can just explain that a bit and maybe this is a good place to sort of set the scene that we need to continue for next time.

Becky Brook

Sounds great. So you're absolutely right. Think depending on what side of the fence you're on, there's been lots of interesting outcomes. So there was a high profile case between Anthropic and essentially Anthropic has prevailed in this particular example for book publishers. And similarly, I'll say similarly, at least there was a meta lawsuit another lawsuit with authors suing Metta for training their model. Both of these came out two days apart earlier this week. I guess actually, I mean, I can very, very briefly summarize what I understand as a non-lawyer the findings are. But actually, if I'm honest, what it really shows us is what I said earlier, which is actually things are getting murkier, not clearer in the short term. That's often the case with these kind of lawsuits where the... where there's widespread infringement. So judges find different things and only get some decisive position from the Supreme Court. But essentially, one of the judges, the anthropic judge, ruled in anthropics favor because they said that it is a transformative use. What I mentioned earlier, it's like Google Books, they're saying that anthropic was using it in such a transformative manner that it was fair use and therefore no license was needed. So anthropic one. Now, meta also won, but fascinatingly, the judge said very different things. So the judge in the meta lawsuit was saying essentially the meta had won, but their commentary was basically suggesting It is not a transformative use. And so what this really shows us is that we're probably going to end up having to go to a Supreme Court in the US at least to have guidance on this. We're probably two years to three years from having real clarity unless there is some political intervention between now and then.

Becky Brook

And that's probably true in most Western markets.

Jakob Wredstrøm

Incredibly interesting. Becky, thank you for being on.

Becky Brook

Thank you for having me. STOP HERE ………………………………………………………………………………………….

Jakob Wredstrøm

That was a great tale. I love that. Even though I didn't come in there.

Becky Brook

I rather could have gone into lots more detail, but I kept it kind of high level because...

Jakob Wredstrøm

Yeah, but I think that was good and also I didn't feel like commenting on it because that was a good cliffhanger. Honestly, that was a super good way to end it.

Becky Brook

It occurs to me, by the way, that Dan, Daniel, Dan thing. I think for end users that might be a bit confusing.

Jakob Wredstrøm

I hope I painted the picture of the bleakness and the optimism and also I explained where I was an inspector before it so I actually don't think it will be that confusing I tried to...

Becky Brook

Yes.

Becky Brook

I guess they've both been on the podcast. So I suppose users.