Victoria Oakley explains how technology, AI copyright and streaming fraud are reshaping the business of music and changing how artists create, distribute and earn.

Streaming has transformed the global music industry. Listeners can discover songs from almost anywhere, artists can reach fans across borders and record labels can build new revenue streams through subscriptions, digital distribution, online concerts and direct fan engagement.

But the same technology is also creating new risks.

What happens when generative AI companies train music models on copyrighted songs without permission? Who gets paid when bots create millions of fake streams? And are Spotify, YouTube and other streaming platforms doing enough to protect artists and their royalties?

In this episode of The Media Room, Vanita Kohli-Khandekar speaks with Victoria Oakley, Chief Executive Officer of the International Federation of the Phonographic Industry, or IFPI. It represents more than 8,000 record labels globally, from independent companies to Universal Music, Sony Music and Warner Music.

Oakley explains how technology is changing the music business through subscription streaming, global distribution, digital fan experiences and new ways for artists to monetise their work. She also argues that record labels remain central because every platform still depends on original music and human creativity.

The conversation examines one of the biggest debates in music today. Should AI companies be allowed to train their models on copyrighted music without consent or payment?

Oakley argues that companies using songs to train generative AI systems should seek permission, enter licensing agreements and pay the artists and rights holders whose work makes those models possible. She also explains why fully AI-generated music without meaningful human creativity or properly licensed training data should not receive the same copyright protection as music created by people.

Streaming fraud is another major concern.

Fake streams, bots, click farms and streaming farms can artificially increase the number of plays a track receives. Fraudsters can then claim a larger share of the royalty pool even when no real audience has listened to the music. This diverts money away from genuine artists.

Oakley discusses why streaming fraud is difficult to detect and why Spotify, YouTube, music distributors and other digital streaming platforms need stronger safeguards, better verification and greater cooperation.

The episode also looks at the future of India’s music industry. India has a vast audience, a growing creator economy and globally recognised musicians, but much of its music consumption still happens through free, advertising-supported streaming rather than paid subscriptions.

Can India become one of the world’s largest paid music markets? Can technology help Indian artists reach global audiences? And can the music industry embrace AI and innovation without weakening copyright, creativity and artist income?

Watch the full conversation to understand how technology is shaping the business of music and how AI copyright, streaming fraud, music licensing, artist royalties and paid streaming could define its next phase.

The Media Room brings together leaders from media, entertainment, technology and advertising to explain the business forces reshaping what the world watches, reads and listens to.

Subscribe to The Core for conversations on business, media, markets, technology and the global economy.

#MusicBusiness #AICopyright #StreamingFraud

NOTE: This transcript is done by a machine. Human eyes have gone through the script but there might still be errors in some of the text, so please refer to the audio in case you need to clarify any part. If you want to get in touch regarding any feedback, you can drop us a message on feedback@thecore.in.

—

TRANSCRIPT

Vanita Kohli-Khandekar: Hello and welcome to the Media Room. Today I have a very interesting guest. Victoria Oakley is the Managing Director for the International Federation of Phonographic Industries.

Basically, it is the apex body that represents all the recording companies in the world. There are 8000 of them that they represent. But the interesting part is that she has a pulse on the music business and there's a lot of talk about what AI is doing, what streaming fraud is to music.

So I got into all of that with her. Over to Victoria. Can you give my viewers a quick short lesson on what IFPI does and then I'll get into the essential questions I have for you.

Victoria Oakley: I'm very happy to and thank you first of all for having me and for having IFPI here to have this conversation. It's really interesting to be able to chat to you. The first thing I would say is IFPI is a terrible name for a brilliant organisation.

Nobody says phonographic anymore. I know you know young people who know what a phonogram is, but we are the global body representing recorded music and record labels around the world. We have 8000 members, so some of our record labels are very, very large.

Some of them are very, very small. Some of them operate in one country only. Some of them operate in one region only.

Some of them operate across multiple territories. So we have what you call very small independent record labels, larger independent record labels. We have the so-called three global majors, Universal, Sony and Warner.

So it's the full range of record labels. And what do we mean by record label? I mean a business that signs artists and helps them bring their music to the world.

Vanita Kohli-Khandekar: You know there's a new, it's been there for some time, but there's a lot of thinking about how the music economy is not just about soundtrack and record labels. It's artist-orientated. I know a lot of jargon comes out time and again, but Goldman Sachs has done a whole sort of metric thing around it and said that the music industry, if you include publishing revenues, if you include live concerts, and live concerts is a big, big element there, is like 105 billion and IFPI number is about 32 billion globally.

And that's also a fair enough number, but how does IFPI view that sort of, you know, and the fact is I think globally because of streaming, because of social media, because of a whole lot of changes that we've seen, there is an ecosystem effect at play. And the IFPI no longer seems like this apex body. There's nothing wrong with that, but how do you view your role in this changing world where platforms are there and the labels are like one cog in the wheel there?

Victoria Oakley: Yeah, yeah, I hear you very much. So look, the IFPI has never tried to be the organisation or an organisation that measures the kind of value or impact of music as a whole. Partly because I would argue as a music fan that's an impossible thing to measure and put a number on.

But as you say, there is revenue that comes in from recorded music being played, purchased through streaming subscriptions, through still buying a CD, which is not dead yet. And we can talk about that a little bit at some point. There's a real resurgence in vinyl purchases, for example.

Of course, there's the whole industry around live music. There's a whole industry around merch, let's be honest. I don't know if any of you have young people who belong to you, but the young people who belong to me, when they go to a concert, the big thing as part of the concert is to queue up and buy the t-shirt, the scarf, the sticker.

And it's terribly important to them to have bought it at the concert. And then, of course, there is this whole growing industry yet to come, I think, about closeness and relationship to your artist that may come in many different forms of engagement. So you're right, revenues from music all come in lots and lots of different forms.

At IFPI, what we do is we look after and measure and analyse specifically recorded music. Now, what did that used to mean? It used to mean an artist makes a record, puts out a record, how many people buy it, right?

It is, as you say, much more complex than that now, but we measure the revenue that comes from that recording. And that increasingly is coming, of course, from subscription streaming. And, you know, we are now at the point where globally we have over 750 million individual subscriptions out in the world.

We're seeing that number grow in every single region of the world. And of course, we're seeing it grow, I would say, at long last in India. So certainly the bulk of the revenue that we're seeing, the bulk of the way in which people experience recorded music is coming through streaming and increasingly through subscription streaming, but it's by no means the only way.

So we are also measuring physical sales. We still measure things like downloads. I know that feels like a long time ago, but, you know, there's still a bit of it about.

We measure what we call sync, which is, of course, when existing recorded music is used either in advertising or TV or film or for a soundtrack or wherever it might be. And we measure public performance rights, which is all to do with, of course, when a business is playing music, whether that is for an event or in the background in a casino or at a restaurant or, you know, anywhere you might go on an airline or somewhere like that. Of course, there's a licencing arrangement around that that ensures that some revenue returns to the artist through that route as well.

So we are measuring that piece. I think that's important to know and understand because, for example, live going to concerts, which I love, and we can talk about that more, I'm going to two next week, or indeed merchandise, they're quite separate businesses and we would argue, or I would argue certainly, the music comes first, right? You record the song, you record the album, the artist puts down their creativity and then puts it out in the world.

And all of those other things like live or merch or indeed some very futuristic, cool engagement online with an avatar with your artist, etc. Those are all things that flow from that first piece, which I think is always the artist recording their music.

Vanita Kohli-Khandekar: It's not fair to use the word power, but a lot of the glamour was with the music companies, let's say, the fun part. Now it is the streaming companies which seem like they are sort of the movers and shakers. And I always wonder how that shifting of that power equation, how recording companies have taken to...

I mean, they're all doing a fabulous job. When you mentioned new revenues, I remember reading this piece about fan parties and you do unreleased songs with fans and you could do meet and greet with an artist in a sort of select listening thing. So there's so many ways.

Somebody was telling me, the Warner Studio head in India and the Warner Music head in India, that Diljit Dosanjh could do a live concert online and for five minutes and you pay whatever, five rupees, ten rupees, just to watch that five minutes. So you could be slicing and dicing this any which way because of the technology on offer. It's a bit like 25 years back when MP3 came and compressed everything and made life difficult.

Just wonder how you look at your position in the music ecosystem as a recording because you are critical to the ecosystem. There's no doubt about that.

Victoria Oakley: Yeah, I think we're more than critical. I'm not sure if you can be more than critical, whatever the right word would be, because everything you've just described, I agree, is really exciting and they are new routes and new channels for artists to put their music out and reach their fans and connect with their fans and on the inverse, new routes for fans to hear that music, engage with it and engage with the artist that they love and whose careers that they follow. But they are that, they're channels, they're routes, they're media in the true sense of the word media, if you go back to its original definition, right? What do all of those platforms or channels or routes need and what do they not have?

Content. It's content. Content is king, right?

And that's what artists are creating and that's what record labels can help artists bring to life. So it all has to start. I know it's very easy for me to say, but Spotify would not exist if musicians did not exist, right?

It is a route and it has utterly changed the way we access music, the way we find music, the way we discover new music, the way we experience it. I think we often forget how utterly radical it is that, you know, in this thing that I'm holding here, I can access music from anywhere in the world at any moment very easily at, let's be honest, pretty low price, lower in some parts of the world than others and I'm sure we'll talk about that, but we have gone from a world where I had to go and spend in my youth, whatever it was, 10 pounds or something, which is a lot of money to me at my youth, to buy a physical CD. I had to wait for it to come out. I had to wait to be able to go to the shop and buy it.

I had to get there before everyone else to ensure they didn't, you know, run out on release day and so on. Then I had to come home in order to be able to listen to it and play it and nowadays I just, I could, you know, open my phone now and say, I'd really like to find some unusual Indian music from, in Tamil and it would be there at my fingertips instantly. So, we have had such a radical change in the way we experience music and exactly as you say, that means that the ecosystem and the plumbing around the business of music looks different than it used to.

What doesn't look different is you start with the music. You start with human creativity and content is the critical thing. So, the roots, the channels, the ways in which we experience it are different.

I therefore slightly question your argument about power and glamour. That's the first I would say. Certainly, the job that I do, there's nothing glamorous about it.

So, I'm not yet committed, you know, when I joined the music industry, I thought, oh, this is going to be so rock and roll. My friends are all so jealous and actually spent an awful lot of time doing really hard advocacy and storytelling and trying to make sure that the business of music is understood by those who are having an impact on it in order that artists can continue to create and fans can continue to experience music. So, you know, it's not quite as glamorous as I thought.

But has the power dynamic shifted? I don't know. I think for me, the power should always be with the artist and the creator first, but you're right.

We are slicing and dicing and seeing new, you know, sources of revenue and indeed, let's say the dreaded words, artificial intelligence, generative AI. I'm coming to that, yeah. And yet another way for fans to experience music and, you know, we have to keep evolving in the way that music reaches people.

Vanita Kohli-Khandekar: I'm going to AI, but I want to pick on that thing you said about looking for Tamil music in India. I can't see you looking for Tamil music in India. Nevertheless, just some quick sense of the India dashboard, because we've lagged behind on the world is at what 837 million subscription last I looked at your numbers and India is like 14 million.

I'm using locally sourced numbers, but we have and we've grown slower than China did in X number of years. You had showed me that comparison. And everybody I've spoken to are just on a large piece of work on the Indian music business as is talking about this hockey stick moment.

What is your sense of where the Indian market is vis-a-vis the rest of the world?

Victoria Oakley: So I agree with everything you said. I think it's really a shame that the Indian music market hasn't grown at a faster pace and hasn't, it's not so much about keeping pace with China, but I guess China is proof that this sort of accelerated growth and sustained growth can really happen at pace. And there is no reason why India shouldn't follow that path as well.

So I think it's a shame and I think it's a missed opportunity. Obviously, the shift that is coming now in India that has taken a while is that shift to paid subscription streaming, which hasn't been the tradition in India for a very long time. But we are now starting to see that.

As I understand it, we are still a very long way off. There is unmet potential, of course, but we are seeing subscription streaming grow. I think India has one of the lowest shares, though, in terms of streaming consumption.

The vast majority of it is coming from the free tiers, the advertising funded tiers, because, of course, these things are never entirely free. Now, what we have seen is a return to growth. So last year, we saw the Indian music market, in terms of revenue for recorded music, grow by 4.6 percent, which is a nice, healthy number and really good to see. The year prior, it had been 2.7. The year prior, I think it had not grown at all. So the good news is we may be exactly as you describe at the upward part of your hockey stick trajectory. And again, you've been talking to a lot of people around the industry, as have I.

There is a lot of optimism that we are indeed in that phase now. I think the other reason that there is optimism is because there is enormous talent in India. And again, you will be much closer to this than me.

You will know much more about it. And people like the head of Warner Music, who you mentioned there, are really investing in finding that talent and making sure it can be brought to the fore. So the numbers are good.

It's an upward trajectory. We had a number of years where I think that growth wasn't there and it looked worrying. And a lot of us were kind of scratching our heads and saying, this is an enormous country with an enormous appetite for music and enormous musical talent.

When is this shift to growth going to come? The numbers though now do suggest, as I said, low growth for a number of years, 2.7 the year before last, 4.6% growth in this past year. And we will reveal the 2026 numbers at the beginning of 2027.

But so far, what I can see internally is that trajectory is heading in the right direction.

Vanita Kohli-Khandekar: Super. You know, last time we met, we talked about AI and the use of AI in the creation of music. And you want to take my listeners through what was the thinking on the use of AI and where the FBI is in terms of lobbying for the right things, if you could just take my listeners to that.

Victoria Oakley: Absolutely. Look, let me start by saying I think we should really distinguish between AI tools, which can be brilliant for artists, brilliant in the creative process. I think particularly true for songwriters, or if you're trying to kind of jam with some other musicians in a particular way or try something out, what would it be like if we had this kind of saxophone instead?

You can't really fly in a saxophonist and practise. In that space, the kind of cool things that musicians in the studio can do with AI tools is absolutely amazing. And I both believe and hope that that will grow.

The other piece is around generative AI. And what we're talking about there is AI music models, a little bit like large language models, but what they're ingesting is vast amounts of music to train the model, and then to pop something else out at the other end. So the output is tracks that have been trained on an existing body of music.

And they're getting better, right? This is what's quite interesting is some of the stuff that's coming out, I hold my hands up, it's not terrible. Some of it is awful, let's also be honest.

So what's happening there is you have vast amounts of content from human artists being ingested into the machine, let's call it, training the machine and the model in order that it can then, when prompted, create tracks. So you might be able to say to one of these AI music tools, make me a song that sounds like Taylor Swift, when she's tired, singing in Tamil, to go back to my previous example. I mean, I'm not quite sure what mashup I've just invented there, that sounds a bit terrible.

But this is the point is you can sort of play with these tools and ask it to make slightly ridiculous things. Now, on the one hand, maybe that's a novelty and it's a bit of fun, right? Like, you know, make me sound like this artist or whatever.

That's kind of a novelty factor and maybe a bit of a laugh. I think where it becomes an issue for my industry is eventually, if you train these models for long enough on really high quality content and lots of it, they will be able to produce outputs that get better and better. They will produce tracks.

And if those tracks then emerge onto the music marketplace and are available on Spotify, etc., you know, for others to listen to, they are competing directly with the input that went into the model to train it in the first place. And so this is where I feel very, very strongly, as do my members, that if a model wants to use our artist's content to train their model, then they need to enter into a licencing arrangement and purchase the right to use that, right? In exactly the same way for me as when I was talking earlier about public performance.

So if a restaurant who's hosting an event or a wedding or a party wants to play music in the background, that's not free, right? In some way, revenue needs to flow back to the artist and then comes to a licencing model. The other place a licencing model works beautifully, let's be honest, is streaming.

So once again, back to my phone, whichever app it is that you prefer to use or whatever platform is the one of your choice, the way that artist music is available on those platforms is because the platforms have entered into a licence with the record label and they're paying money to have it on their platform. So we believe exactly the same principle should apply. If an AI model wants to use an artist's music, they need to seek permission.

They need to enter into a licencing arrangement and they need to pay. I don't think this is a revolution. Just as I've said, you would expect to pay for it if you were playing it in the restaurant.

You would expect to pay for it if it's a streaming platform that's buying and using the music and making revenue from it. Exactly the same way if an AI model wants to use an artist's work, seek permission, enter into a licencing arrangement and pay appropriately. The risk is that in the absence of that, and there are a number of parts of the world where this is happening and a number of infringing models that are going ahead and doing this, they're just absorbing content.

They're ingesting it free for all. An artist neither knows whether or not their work has been used nor has been paid and compensated for it being used. And as I said, then the model can pop something out at the other end which could compete directly with the original tracks.

This sort of free for all, I think is very, very dangerous. What you will sometimes hear from some players on the tech side in particular and those who invest in these innovative AI models is, oh, we can't possibly be expected to go and find the rights holder and check with them and enter into a licence for every single piece of content we want to put into our machine and train on it. It's all terribly difficult, too much for us.

I'm calling it out. That is a rubbish argument for two reasons. One is licencing happens.

We make it really, really easy at record labels. We want to licence. Come to us and say, I would like the body of X work that Universal Music or Warner or Sony or whoever it's the right solder for.

We will enter into a licencing agreement with you. Job done. You pay.

There's a deal. It all works. The idea that it's difficult to do, I think, is unfair.

The second reason I think it's absolutely rubbish to say we can't possibly go and find, for every single song, find the rights holder. It's too much, too difficult for us. We don't know.

We can't possibly know what's going in there. I tell you, having worked on the tech side for a number of years, if you are building a good model and you believe in your model and you're a big company that is then going to make money from what you put out coming out of your model, you know exactly what you're putting into it. A good chef knows exactly what their produce is, where it came from, what ingredients they're putting in, what they're serving up to their customers.

They know for sure what they're putting in.

Vanita Kohli-Khandekar: This is an argument I use in publishing as well because ultimately, there's no incentive for me to produce original content or write original stories. If you're going to finish it all off, then what will you do? You'll be hallucinating 50 years down the line because you'll have no original stuff left.

But no, we are digressing. I wanted to ask you, you mentioned about, you know, we talked about the use of AI and you talked about these distinctions, but the copyrightability of, sorry, is that a word? Is that music or whatever is created copyrightable and there you had drawn a distinction and I talked about the lobbying efforts.

If you could just take my ears to that.

Victoria Oakley: Absolutely. So, we are very clear that a piece of music, and I put that in inverted commas quite deliberately, that has been created solely through generative AI with zero human input and in particular, which has not sought a licencing agreement with the rights holders for the music it trained on, any output from that is absolutely not copyrightable. It should not benefit from copyright protection.

It's not what copyright laws were designed to protect and indeed, it's come about through an abuse of copyright. So, it would be absolutely perverse in my mind to say, you can abuse copyright to make your thing and then when you've got your thing, you want to seek copyright protection for it. It's kind of, you know, it messes with my head to think about that concept.

We are very clear that entirely AI generated music that didn't come through a licencing and ethical arrangement and that didn't have any human involvement should not benefit from copyright. You are increasingly seeing that being recognised and acted on by some of the platforms out there. So, I would say even in the weeks and months since you and I had the chance to meet in person when I was in India.

Actually, we weren't in person. We had to have a video call anyway. Even in the last weeks and months, we have seen a number of platforms and you can look at something like Deezer, the very well-known French platform or even some of the things that Spotify is doing.

They're coming forward and saying, we've got to find some ways to get this right. We don't want to accidentally be paying out royalties on music that isn't real music and doesn't benefit from copyright protection. We as platforms are trying to find ways to make sure we're putting in place the right sort of filters to ensure that what reaches our audience is appropriate and correct, but also what receives payout is appropriate and correct.

Because those are two different things and there may well be an argument that there should be space for AI generated music if licenced and ethically trained to exist and to be played. But that's different to the question of whether or not it should receive royalties and a share of the revenue, which is where the copyright ability element comes in. So, in terms of lobbying, we are very clear about this and we are talking to governments all over the world, including when I was most recently in India where we were talking to the government there.

What would I say? I think this is a really difficult circle to square for governments all over the world, and we are seeing different approaches. What they want is to try and make sure that innovation can happen, that growth can happen.

Let's be honest, that tech firms, including those who build AI models and AI models for music, want to come to their countries, invest in their countries, build data centres in their countries, hire talent in their countries, and that's what tech firms are promising left, right and centre. So, I understand that governments really want to make that happen and make that the case. I worked in government for a long time and I understand that protecting your citizens and building an environment where growth is at the forefront is the number one duty of governments.

What I think they are doing, though, is they are slightly listening to the promises of tech firms and the argument from tech firms that they need this free-for-all and they must have access to everything. And I think there are very strong arguments why, in something like medical research or road planning or how we design systems to ensure that anti-malaria drugs can be properly distributed across a network, in all of those sorts of situations, it's absolutely critical that we are using AI quickly to better address those societal problems. And in order to do that well, clearly AI models will need access to vast amounts of data and content in order to train well and deliver solutions in that space.

And there, I think there is a genuine debate to be had that is not my debate but is about whether or not we can make access cheaper, easier, quicker for those sorts of purposes for AI models. Where I struggle is when somebody says to me, in order to help solve that problem, you know, hospital bed space planning across a giant network of hospitals, for example, or, as I said, infrastructure design or medical research where simply AI can do it so much quicker and better than humans can, that in order to do all of that, what AI models need access to quickly and freely and without either paying for it or seeking consent is all of Taylor Swift's music, right? And somebody needs to explain to me how those two things come together, right?

Vanita Kohli-Khandekar: Totally, totally. I am so much on board with you on this one because I've written so much about this. But taking this further and we've discussed this fake plays, where is IFPI on fake plays and how badly are they impacting the music business?

Victoria Oakley: Yeah, well, this is another very serious issue and it's an issue that existed before generative AI came along but I think it's being sort of supercharged by AI in a way that is making this problem easier. So we call this streaming fraud and I'm very intentional about calling it streaming fraud because it is fraud, right? It is a form of theft.

So what we're talking about here is fake plays. So what that means is a track and that track might be real or AI generated, you know, in either instance. But what happens is you find armies of bots or streaming farms or click farms and you have them falsely spin up the number of times it looks like a track has been listened to.

So in reality, a track may have been listened to either not at all or just a sort of moderate number of times, often sitting in the sort of middle ranking of tracks on a platform. So not the sort of uber top artists but nor the stuff absolutely at the bottom either. So it's a kind of invisible space for these fraudsters to hang out.

So what they do is they take a track, as I said, falsely spin up the number of plays. That means without it actually having been listened to by real fans, it looks as though they merit a greater or indeed, you know, possibly went from non-existent to large proportion or share of the revenue that comes through that subscription arrangement. So what you've got is, you know, a fixed pool of money available that is shared out across artists according to how much they are listened to or played on a given platform, all of which is commercial arrangements dealt with between record labels and the platforms and not by us.

But you know, the way it works is there's a pie and it's shared out according to listenership, artist by artist. If in the midst of that pie, you've got a chunk of fake artists or fake tracks but certainly with a fake number of listens, then they are taking a chunk of that money for themselves, even though that music wasn't actually listened to that many times or indeed at all, which means that money isn't available to go to the artists who were actually listened to. So I think it's pretty outrageous.

I think it's what I call very lazy fraud because you don't have to do very much, right? You kind of sit there, you make some fake tracks, you hide it, you know, you move it through multiple kind of VPNs and I don't know what, and then you have a bot and a click farm or in some instances, as I understand, even actual humans who are paid to just sit there and listen to a fake track over and over again, just, you know, press the button, press the button, press the button. And I would say a significant amount of the revenue that should be going to the artists who've truly been listened to is going then to these fraudsters.

It's very hard to track, it's very hard to find, it's certainly hard for us as IFPI or as record labels to see exactly where this is coming from. And for me, the obligation on who can fix this sits very squarely with the distributors who are the kind of gateway to how music gets onto a platform and also with the DSPs themselves who need to be doing much more to see where this is happening, to track it, to have really good know your customer tools in place. And then the thing I'd really like to see is, I think where you do have progress and some distributors or some platforms are working on this and trying to put tools in place, we need to see a much better system for sharing, right?

So that a fraudster can't get shut down on one platform or through one distributor and then just jump to another one and do it over there instead. And then we're playing this game of whack-a-mole. So I would really like to see that side of the industry kind of find ways to come together to stop this collectively.

It's a real shame. And we have fairly strong indications that the people carrying out this fraud, who are generating these fake plays and therefore this money that is an unfair and incorrect share of the revenue pie that should have gone to other artists, they're doing it often associated with other crime. So it's very rare that you have a streaming fraudster who's just sitting at home carrying out streaming fraud.

It's usually part of some sort of broader organised crime network often. It's often being used to fund something else. So I think there's a real case as well for law enforcement all over the world to start looking at this much more seriously.

Vanita Kohli-Khandekar: How do a Spotify or a YouTube or any of the platforms, how do they look at it? Do they have safeguards in place? And before you answer that, what is a DSP?

Victoria Oakley: A platform, a platform, a digital streaming platform. So any of Spotify or whoever is DSP. Apologies for the shorthand.

I think those platforms, YouTube, Spotify and all of the others, I think they're looking at this hard because they do know that this is a problem and they don't want to be disappointing their own customers by paying out revenue to fraudsters rather than to authentic artists. Are they doing enough? I'm not convinced yet.

Certainly we are running a small intelligence operation here out of IFPI where we are still finding fairly serious cases of fraud that is happening everywhere. There is no platform who is immune to this. So by definition, if we're still finding fraud, then not everybody is doing enough.

Vanita Kohli-Khandekar: Any estimates on how bad it is? I mean, if you talk about a $32 billion recorded music industry, how bad?

Victoria Oakley: It's very hard to measure because again, you know, you have to have a really intricate system. And what the way that we do that is we try to essentially have what we call sort of anomalies or flags. So the sorts of things that might indicate that streaming fraud is happening rather than real listenership.

And so those things might be something as simple as do all of the plays come in one big period, the same track being listened to over and over again, literally one after the other on one device in one place, for example. Then obviously it would suggest to you there's a series of flags there that mean this is worth looking at. So we are doing work like that.

What we're not able to do is say it's this bigger problem or it's that smaller problem. I would say in a way, whether it's 1% or 10%, it's still bad. And as you say, we know that it's a $32 billion industry.

So you can do the math and say that is how much money is slipping away from authentic artists to whom it should be going and going instead to criminals. We have seen some reports in the past, the French government did an investigation that suggested it might be as high as 10%. There's a very interesting tech music organisation called BeatDap.

They looked at this issue a few years ago. They thought it was around 7%. Now, these are relatively unreliable measures, but it gives you an indication that this is not nothing.

Vanita Kohli-Khandekar: That's interesting and I don't know how they will tackle it. One last question, given this new shape of the business, if I was to ask you as a head of IFPI, what are the two, three big challenges or challenges is not a negative word in the sense, or what are the two, three big things that occupy Miss Oakley? What are they?

Victoria Oakley: Well, we've talked about AI and that has to be top of my list because as I said, if we can get this right and put the right guardrails in place that have respect for content, then I think this new advent of generative AI can be a really interesting new source of revenue for artists and also a new and interesting experience for fans, but we have to get it right. We really do need to put the guardrails around the way that generative AI uses music. That is absolutely the number one challenge.

I think my second one is streaming fraud, which you just asked me about. Any business that is looking at a problem like that where money is seeping away to criminals who are able to, quite frankly, steal revenue that should have gone to true creators and without being true creators themselves, simply kind of seep that money away to themselves, we have to find a way to fix that. I've got guys working hard with distributors and platforms to really push them to put in place a number of measures.

We think it's not difficult to fix. We think it requires some intentionality around deciding to fix it, but I think for me, perhaps the one that I put third, even though in many ways it's the most important, is that we must keep making sure that record labels are doing the thing that record labels should be for, which is taking an artist, taking talent, providing a whole host of professional services that then sit around that artist and that talent to ensure that that artist's work can reach fans. Exactly where we started this conversation, which is that looks different now than it did 18 months ago, than it did 20 years ago, than it did 50 years ago. It's going to look different again 18 months from now and heaven knows what it's going to look like 20 or 50 years from now.

So record labels have to keep making sure that we are innovating in the ways that people want to experience music so that we can properly partner with artists and be the kind of business support that they need to get their work out into the world.

Vanita Kohli-Khandekar: You know, it struck me and it's something which is so axiomatic to our thinking now in the world we live in, that streaming has created one large global market from the demand and the supply side. Anybody can come on and showcase their talent and anybody can be supplying or distributing. I mean you could be reaching to most markets.

So it has created homogeneity of access and you know you can sample, but therefore the business opportunity, you know one of the things, results of it for example, is the collabs I see between Indian singers and so Ed Sheeran and Arijit Singh or you know Adiljit Dosanjh, and I love them you know. I mean I'm talking about the Indian, I'm sure there are collabs between Spanish artists and American one. You know, I think it's such a fabulous opportunity.

I don't know if the IFPR has any role or any point of view there, but I was just curious how you look at it.

Victoria Oakley: I mean we've spent a lot of time looking at the data that we have, that looks at that I guess export story right. So you know there are plenty of criticisms of the streaming model and some of them are fair right, but it has done an enormous number of things around access and democratisation of music and I think this point that you make about opening all of us up to new worlds of music and new sources and therefore creating that demand right, that we probably didn't even know was there for these kind of collaborations or mashups.

One of the things I'm fascinated by is there's a record label whose name I have forgotten but it will come back to me, that is now based in a part of Canada that specifically signs Indian artists because of the Indian diaspora in that part of Canada. I mean isn't that brilliant that there are people who live there who are both Indian and Canadian, they want to listen to music that is somehow a little bit of both because they are a little bit of both and now have a record label finding artists to bring to exactly that market and I don't think we would have had any of that without this sort of open access to the world and I think at the same time though what I'm seeing my record labels do is invest more and more locally. So whilst you've got these global deals and this wide access and this kind of I can open my phone now and type in anything and find music from all over the world, at the same time you're really seeing record labels spend time in country, on the ground to truly find local talent. What they can do though now is they can sell that local talent well beyond the local place that they're in.

Vanita Kohli-Khandekar: Correct, absolutely. That's exactly what this Warner guy was telling me about but it's very interesting and you see that on the video streaming site too. Lovely, this was a lovely conversation Victoria.

Thank you so much and I'm waiting to hear which is your favourite Tamil track. We shall, we shall.

Victoria Oakley: I'll get back to you.

Vanita Kohli-Khandekar: I can die for you. But lovely, thank you so much for talking to us.

Victoria Oakley: Thank you for having me.