Who Owns the Future of Music?

Jennifer Strong (00:00): So in a world where AI can generate a radio ready track in a few seconds, what's the human element here that AI doesn't really touch? Andrea Czapary Martin (00:09): Well, I think, you know, the question is who owns the future of music? We represent 180,000 composers, writers. We license the music wherever it is played around the world. It's about a $1.6 billion business is definitely not the models. Definitely not the models. I still think the human aspect and the emotion that come from composers, writer and artists will stay. So I think that for me to answer that question is quite clear. I think it's the people that prove who wrote the songs and who get up on stage and sing it are the ones who own the music. Ghazi Shami (00:52): We had a hit record last year in the United States called Rubbers that did, I think close to 100 million plays now. And it was primarily generated with AI. So it was at the center of a lot of conversation on Instagram and TikTok and a lot of blog posts. There was a significant argument on whether it should be nominated for a Grammy or not nominated for a Grammy. Um, how much human interaction has to happen for it to be nominated versus not nominated. There was definitely human interaction in the record. It was recut, it was polished. Some of the lyrics were changed and so on and so forth. But there was a, it was at the center of conversation. And I think being at the onset of where the business is headed is, for me, an, an exciting to be at the center of it and helping navigate what the human construct is gonna be going forward and how we manage that and protect it. Jennifer Strong (01:42): Yeah. No, absolutely. Now, Andrea, you alluded a bit to this. You collect over a billion pounds annually for songwriters. And you've got clear red lines around things like copyright. And how much has your world changed recently? What does it mean now? Andrea Czapary Martin (01:58): So of course we spend a lot of time lobbying and making sure that the copyright is protected. Um, but we also use a lot of the AI to find sources of revenue that our owners, which are the members, um, should be getting. There's so much money that are not paid to composers and writers around the world, and we're using AI to find those sources of revenue, which are much stronger than what we currently have, you know, very, very modern systems to collect that money and to a- attribute it to the right people. So, you know, our, as much as we talk a lot about, oh, AI is gonna hurt the music industry, there's a lot of AI music that you listen to and you don't even know it's AI music and, you know, it's most of it is free. I say that AI is also very good. Andrea Czapary Martin (02:55): You know, you heard about Gazi, you know, using AI to make music, but there's still a human touch. Also, AI to find mus - you know, the music and the money that is out there. So for example, you know, the music industry is a little bit archaic, much more than I ever thought before I became CEO seven years ago. When you listen to live music, you go to a concert, uh, Sazam cannot recognize what is the playlist. So a lot of the playlist is actually all mainly manual. So we used AI to scrape the web around the world because most of our, our members are, you know, worldwide. Um, you know, we're biggest, second biggest exporter of music. And that allowed us to be able to find out what are the music that's played, who should be paid. Um, and so that is a very good example of AI that is good, uh, because it's more money. Jennifer Strong (03:51): All right. Gazi, you built Empire betting on humans when the labels otherwise were trying to look for safety. And I'm wondering for you now, how much has changed and how you, like is this more in this moment of things being flooded by AI without humans, to your point? Um, is it a marketing problem trying to help the human artists break through? Is it a pipeline distribution problem? Walk us Ghazi Shami (04:16): Through. Um, well, I was born and raised in San Francisco where all music technology and convergence of music and tech started. So we've always been a tech-forward company. A lot of people don't know that Empire is as much a software company as it is a record label. So all the supply chain systems, the distribution, the FinTech payments, like all that. So in a nutshell, um, I don't see it as a problem. I see it as an evolution of where humanity and where society is headed as a whole. And how do you adapt to that evolution? Mm-hmm. And so as a company, because we've always been a software company, the adaptation for us has been a lot more natural than maybe some of these archaic, um, companies or systems that are outdated in the music business. Our application for AI was pretty intentional because we were already doing a lot of like machine learning - uh-huh. Ghazi Shami (05:15): Seven, eight years ago, and then that naturally transitioned into more agentic programming, whether that's, um, you know, the analytic systems that we're building that are doing predictive modeling for records that are growing faster in the system, whether it's our legal team uses Spellbook to draft faster and to redline contracts a lot faster and to be able to call things back up. Um, we're building a whole entire layer where we're using the API from DocuSign to feed into our system and then we can search contracts faster and pull terms a lot faster and then just prompt. Um, in the studio from the creative side, which is where I grew up, we're doing a lot of STEM separation for Atmos to be able to take old music catalogs and redeliver them in Atmos on music services. Um, we're replaying a lot of samples, um, that are hard for people to replay because you can't access the musicians or bring the musicians in. Ghazi Shami (06:07): So now you can supplement some of the session players with, uh, systems like Google Flow or Suno or Udio or Trebelo. Um, I think the, the primary construct is as we grow, uh, as a company and as an industry to make sure that we instill the guardrails to keep everybody pointed at a north star that's ethical. Um, and, and just like any other tool that comes, there's always gonna be bad actors and good actors and you just wanna make sure that the Googles of the world and the Spotifys and the record labels of the world and the publishers and the PROs and the societies and so on and so forth are advocating for the good actors and setting up guardrails for people to, you know, be incentivized to do good with AI. Jennifer Strong (06:53): Of course. And we should also say you're a little humble up here. He writes a lot of his own software. You always have, right, for the company. You've been quite active. Ghazi Shami (07:02): I've, I spend a lot of time in the studio. Yeah. It's my first love. I'm like a kid in a candy store when I'm in the studio. But, um, it's, it's been pretty remarkable to watch the acceleration on production from a creative standpoint - uh-huh. Having access to AI. Because when I was a kid, one of the primary things that precluded me from continuing as a producer was I couldn't afford all the session players. I couldn't afford to hire the guitarist, the bassist, the drummers. Now, if you're a creative and you're a producer and you have a creative idea in your mind, you can create all those stems. Um, and you can write all those guitar licks and notes that were in your mind that you couldn't afford to hire those people with. Right. So I always, you know, I have a mantra in life that access creates culture and I think AI, uh, creates a level of access to creatives - uh-huh. Ghazi Shami (07:52): And puts something in the, like a magic wand in their hand that they've never had before. Look, at the end of the day, um, music is about taste and just because everybody has access to the tools doesn't mean that they have the taste to create something that humans want to consume at a high level. So, um, I think it just accelerates the creative process in ways that we've never seen before. And to me, I find that pretty remarkable and exciting. Jennifer Strong (08:17): Absolutely. Andrea, I would love to hear more about what you're doing with machine learning. And I'd also, at some point, both of you would be great if you could brag a little bit about the work you're doing and also tell us how it is you're gonna fight for AI for good and build those guardrails. Andrea Czapary Martin (08:29): Yeah. I mean, the, the, you know, the music industry, like I said, as much as we think of all these artists, it looks very modern. Um, the way it's managed is, is very complex. So PRS for music, we process three times more data than Visa transactions in the world, just to give you an idea. Uh, because each song, there's at least six composers and, and writers - uh-huh. For each song and each of them have publishers, labels, et cetera. So you can imagine how complex it is. Of course, AI has helped, and there's a lot of middlemen in the music industry. And one of the reasons why I was very passionate at PRS, and we doubled our revenues, you know, the last seven years, is composers and writers should be making more money. 'Cause without them and without the artists, there's no music. Mm-hmm. Right? And so, um, Andrea Czapary Martin (09:25): We really have to improve on that. Um, and AI is really helping us. Like I said, I talked about set lists, but also we used AI, um, agentic AIs. We built it agents to look at the international revenues because we're the second biggest exporter of music in the world. Like, you could imagine the revenues coming around the world into PRS. Um, and we know very well that the global societies, because almost every country, I would say every country has a society, if not more than just one. The US, for example, there's five of them. Most countries only have one. Um, but they, they are very archaic, um, and, and it's very local. So we had to build some agents to be able to find sources of revenue, and we have. Um, and there, you know, there's a lot of money not coming out different countries, you know, like India and Africa and sometimes a little bit Latin America or the Gulf States. Andrea Czapary Martin (10:27): And it's not just the DSPs, the, you know, digital service providers, but the Netflix, the Amazons of, of this world. So that has helped a lot. Um, and, um, using AI for that. I think there's gonna be a lull though a little bit for our, our, the revenue because, you know, because of all these AI music out there that is, they are not paying anybody. I think what's gonna happen is, you know, with abundance, scarcity is gonna be even more valued. Mm. And, you know, the human side is gonna be even more valued. And we see that, you know, in the last three years, our live revenues for our members have more than doubled because people are going to concerts. Jennifer Strong (11:10): All right. So Ghazi, if a fan-made clone of one of your artists' voices goes viral, what's Empire's Playbook? What do you do? Ghazi Shami (11:19): Well, we were talking about that backstage. Um, you know, one of the onsets of the music business in the streaming era that created a lot of revenue and gave artists a lot of optionality early was when YouTube built something called Content ID. And that was probably like in 2008, 2009, somewhere around there. And what that was, was basically the modern construct that you see now on TikTok and on YouTube shorts where people are making fan-generated content and dancing to a song. Um, a lot of artists weren't fans of that before. And so it allowed artists to either have a block policy or a monetized policy and then the system would claim it. And so now what we're doing, um, we're working very closely with the YouTube team and the DeepMind team at YouTube, uh, or at Google. Um, they're building something called voice ID, which essentially if somebody uses AI to clone somebody's voice, like say Chris Brown or Michael Jackson or whatever, then you can put either a block policy or a monetized policy on that. Ghazi Shami (12:22): Maybe as an artist, you think it's cool when people make - uh-huh. Tracks that sound like you. Maybe as an artist, you don't find that attractive. So you have the ability to block or monetize. And these are the guardrails that I'm talking about is giving the artists choice on what happens with their creative, um, output. And so we're doing that with several other companies as well. Um - Jennifer Strong (12:45): Awesome. And have you faced this yet? Fan produced, you know, AI clones or vocal Ghazi Shami (12:50): Clones? Yeah, it's been going on. We were one of the first companies to encounter it. Um, we had somebody that knew how to download certain language models that had voice cloning. Mm-hmm. And then he would go in the studio and sing the records. But he was a writer. And so he, he would say, "When I send this record to The Weeknd or to Chris Brown or Shabuzi or whoever to try to get a placement, I wanna send it in their voice." So he would go sing the records in their style and then apply the voice model to his voice. And then one day he got a little slick with it and he put it on Apple Music and put it out on YouTube and the industry had a meltdown. Um, but we've come a long way from that. That was two and a half years ago. Ghazi Shami (13:33): So in a short am - you know, you look at the rate of acceleration in technology and a lot of people are not prepared for that. But as human beings, mentally, we've also shrunk that delta down pretty quick as a society. Mm-hmm. And what's acceptable ver - now versus what was acceptable just two years ago. So when you, when I look at my children who are, you know, nine years and younger - uh-huh. Their, their consumption habits and what's okay to them is a lot different than what's okay to their 50-year-old father, you know? So, um, I think we're gonna see a massive change in the next - Yeah. 24 to 36 months. And being somebody who champions the music industry and champions creatives, I wanna make sure I'm the front end, I'm on the front end of that change, because I wanna be able to steer the ship. Ghazi Shami (14:19): So we spend a lot of time working with the YouTubes and the Spotifys and the Apples of the world. A lot of the leadership teams fly into San Francisco, come work out of our studios, watch the user behavior of our producers, our writers, have personal conversations with our artists to see what they're cool with, what they're not cool with. Mm-hmm. Um, and then I, I kinda put myself at the center all the time because as much as I am a technologist, I'm also a creative at heart because that's where I, I started and my, you know, my, my core is still there. So I feel like there's always a happy medium where you can build it to appease people. Um, and at the end of the day, the most important thing is to be able to monetize the creative so that creatives can make a living doing something that they love. Jennifer Strong (15:00): Right. Exactly. All right. We're in our final few minutes. What's the most important thing each of you are working on right now? Andrea Czapary Martin (15:06): Well, I'm in transition because I'm leaving as CEO, but, um, I wanna continue in the music industry. And why I wanna continue in the music industry is to make sure that there's less people taking money out of ecosystem when the music is played when it's paid. Um, and so that's what I wanna make sure is that we reduce that, those different commissions that people get, and also the time. People don't realize when you play a song, it takes almost sometimes up to nine months, 12 months before that composer, writer, and artist is paid. Less live, but when it's on a digital platform - Ghazi Shami (15:48): If they're ever paid. Andrea Czapary Martin (15:49): Hey? Ghazi Shami (15:50): If they're ever paid. Andrea Czapary Martin (15:51): If they're ever paid, right, exactly. Yeah. If they're ever paid. And that I want to, because if you think of it, in a digital world, if you play it, why wouldn't they get paid the next day, right? We know who it is. We know what the split is. We know what, et cetera. So working on, on the lag transparency and also the amount of money they get. Jennifer Strong (16:11): Hmm. How about you, Ghazi? Ghazi Shami (16:14): Um, well, so Empire is a publisher, a distributor, and a record label. And three years ago, I decided to bifurcate the company into a software company and an entertainment company. Um, and so for the last few years, I've been architecting a new system called Supply Chain, which is basically a rights management system for the music business. And I, I believe that there was a massive hole in the music business on how to manage rights - Yeah. In terms of going across the board horizontally from a music documentary all the way to merchandise and then music masters and publishing sitting in between. So if you're a private equity company and you went and bought Bruce Springsteen, you kinda don't know how to monetize it. You're just collecting the revenue streams from wherever they're coming in. Right. And as we increasingly go into a financially driven music model, I wanted to build a rights management system to be able to take care of all that. Ghazi Shami (17:07): So, uh, we're gonna officially announce it next month here in, at ADE in Amsterdam, and we're actually gonna HQ that company here in Amsterdam. Cool. So it's pretty cool for me to be here talking about it now. I didn't expect you to ask me that question, Jennifer Strong (17:20): But - Well, I didn't expect you to say Ghazi Shami (17:22): That. Yeah. So serendipity.