Striking the Right Chord: How Musical AI is Keeping Artists at the Heart of Generative AI
Sean and Matt’s Innovative Solution for Attribution, Licensing and Profit Sharing in AI Music.
Attribution is the golden ticket for a happy marriage between the music industry and generative AI. Without it, artists don’t get paid—kind of a deal-breaker when it comes to industry buy-in. Enter Sean and Matt, seasoned pros in both music and tech, who’ve been around the block enough times to know that building a company and navigating the maze of music rights is no easy gig.
But with their fresh perspective, they’re mapping out a future where AI and music don’t just coexist—they create a symphony of sweet, profitable harmony together. By addressing the critical issue of attribution and fair compensation, Sean and Matt are paving the way for a new era where creativity and technology converge to the benefit of artists, rights holders, and the industry as a whole.
In this interview, Matt and Sean share their experiences in music and tech, discussing the challenges of navigating music rights in the AI era. They offer insights into how AI and music can work together while ensuring artists are properly credited and paid, and outline their vision for a future where both industries thrive side by side.
Some background into Musical AI:
As ardent life-long supporters of the music industry we surveyed the Al music landscape and saw the music ecosystem spiraling into unethical practices, so we fixed it.
With an unshakable goal of ensuring fairness and equity in Al music generation, Musical Al was born. To make it happen, our team of music and tech stars have come together to create an Al generation licensing and attribution platform to uplift musicians, respect rights holders, and unlock creativity for every music maker.
Thanks for taking the time to talk today. Let's start with some background about the founders. What made you want to get into the Music AI space? What experiences led you to founding Musical AI?
MATT: I've been in the music business my whole life and the music tech business longer than most people have been alive. I started working in record stores when I was a kid in the eighties, which I love. Early on, I worked at an important independent record label in Chicago, Wax Trax Records, home to Ministry and side projects from Nine Inch Nails. I had a number one record with the KLF when I was there. Being British, hopefully, you know the KLF; we put it out in the US. Then I left to start a house music label. I was the first straight white boy in Chicago to start an independent house music label, putting out records by Derrick Carter when we were both teenagers.
Then the Internet happened, and I immediately saw how fun it was to see how people were using music. As a DJ, you watch the dance floor; if you clear the dance floor, you're not doing your job. Working in music on the Internet, I had data that showed me exactly what people liked and didn't like, and I was hooked on delivering music that brings joy. I went to work for Motorola, building early music streaming technology. Then I was at a company called MusicNow, the very first streaming service with all the major labels. We sold that to AOL. Then I found my way to Napster in Los Angeles.
I ran product and music at Napster when it was legal when we were paying people. Then I went to Beatport as COO. Six months later, I became CEO, grew it substantially, helped sell it for the investors, and started another company called Metapop, which I sold to Native Instruments and worked there for a few years. After the pandemic, I wanted to work in AI music. I had been consulting for a while, and I knew I wanted to work in AI music. Given my background, most people will take my call; they'll talk to me for five minutes, even if I annoy them.
I talked to a lot of people in AI music, and they were all on the wrong side of history. I've seen how these things play out before. I was talking to a friend, Eric, saying I couldn't find an AI company I thought would succeed in the long run and be supported by the music industry. Then I met Sean. Sean had already put together a team that had worked on all the necessary technology for the health of this ecosystem. I was lucky to have Sean ask me to join the team.
SEAN: We're the ones who are lucky. You never know with co-founder dynamics how things will shake out. It's rare for relationships to work well and be low friction from the beginning all the way through. That's been the case with our team. I've had a lot of co-founders, and all of them have been amazing. This is the best co-founding team I've been a part of.
I come from the infrastructure, AI, and data side. A key to our organization is having people who understand these critical intersections. Without that wealth of knowledge, we wouldn't be able to bring to market what we are. I've been doing startups for 20 years. In 2004, we founded Coradiant. It was a big Canadian success. We had 150 employees by the time we sold for nine figures. None of the founders made any money off that $160 million deal.
We learned about dilution, fundraising, and pitfalls in fundraising from the Valley, which set us up for success in our subsequent startups. I think, at last count, there were a hundred startups born out of employees from Coradiant in Canada. I can checkmark a couple of them. After that, I was part of PostRank, a Waterloo-based startup. We sold that to Google. Then I was with CoTweet in the Valley, which we sold to ExactTarget, later acquired by Salesforce.
I've been doing startups from an infrastructure perspective, mining data and thinking about that world inside and out. AI is heavily infrastructure-based. If you're working on deep AI and deep learning, you need to understand the AI stuff, which our CTO Nicolas Gonzalez Thomas does really well, the infrastructure stuff, and the business stuff, which Matt understands in the entertainment ecosystem.
Our backgrounds led us to found Musical AI and build our vision together. Unlike some CEOs who claim a singular vision, ours is a combination of the founders' views on AI's potential. We all believe in AI, but the current approach doesn't work. It's not sustainable. Companies are using humanity's content and saying whatever comes out of the algorithm is theirs to sell, without concern for what goes into it. We find that fundamentally insane.
As pioneers in music rights management for AI, what are some challenges you've encountered in this emerging field?
SEAN: I think there are a few challenges. One of them is navigating the complexities of existing rights structures that weren't designed with AI in mind, creating a gray area. This ambiguity allows anyone—Jack, Jill, or Mary—to say, 'Oh, well, I think it should be X,' and build their company around that idea without any legislation or laws providing clear direction. I think one of the challenges is really about this type of structure.
Another challenge that comes to mind is the knowledge gap between different groups, organizations, and people who don't fully understand the grayness I just mentioned. When it comes to AI, things aren't strictly black or white—what can be used, what shouldn't be used, and how things like copyrighted material will likely end up being honored.
Those are a few challenges that come to mind.
Matt, what do you think?
MATT: Well, I think you said it really well. The only thing I can add is that it's a rapidly evolving market. There aren't a lot of people making a lot of money, yet there are a few companies making most of the money in AI. And I think, like any emerging technology, it'll be a bumpy road, and we're all going to learn a lot in the coming years. So, being responsive to all that change is one of the challenges that we've already figured out how to manage. We've been at this for over a year and a half now, and we've listened to our customers, which has had a massive impact on our strategy and what we're doing.
I wouldn't call that a challenge for us, but it could be a challenge for a lot of startups if you're not mature enough to put aside what you thought was true when you learn what is actually true.
This is the fifth time I have built a massive database of licensed music for a variety of purposes. Every time, you need to have the master rights cleared from the labels and the clearances from the publishers in order to do streaming, sell a download, or put music in a movie. So, this is not new to me. It is the first thing I explain to new arrivals to the music business from the tech world. I explain there are two copyrights. That segment of this market is going to move slower than the segment where rights holders own both sides.
Currently, in AI production, catalog libraries and indie labels that own both the publishing and master rights can enter this market quickly. This is possible because there isn't yet an industry agreement or government regulation on how to split revenue between the publisher and the master rights holder. We've built a system that's flexible and can adapt to whatever approach is decided in the future. However, I believe this issue still needs to be fully resolved. In the meantime, we've managed to navigate this by partnering with those who control both sides.
Can you talk about some use cases for your attribution technology? What are the potential benefits for artists, producers, and the music industry?
SEAN: When I talk about the industry as a whole, I’m specifically referring to the provider side—the AI companies. These companies, as Sam Altman from OpenAI has mentioned, rely on high-quality copyrighted content. I'm combining these two areas—those creating AI models and the excellent content those models are trained on—because they make up a significant part of the AI industry.
For this industry, our approach to attribution is really about promoting transparency and accountability. It allows rights holders to feel confident working with AI companies without the fear of losing control over their IP. At the same time, it enables AI companies to use copyrighted content in ways that save them a lot of time and give them and their investors confidence that they are engaging positively with the ecosystem.
Matt, I'm not sure if you want to get into the specifics, but I wanted to start by discussing the broader picture—what does the use case for attribution solve? We can go into the specifics if you want. Matt, it’s up to you.
MATT: I'll just give you a simple example of the particular. In the realm of Spotify—and let's put aside arguments about whether Spotify pays enough or not—the way they decide who gets more money or less is based on what music gets listened to more, which as a basic function seems to make sense to me, right? The music people listen to more should make more money. I used to live in a world where the more you sold, the more money you made. But, okay, we'll move to a model where the more people listen, the more money you should make. Imagine if Spotify instead decided, 'We're just going to pay you based on the volume of junk you deliver to us.' Then, all of a sudden, Madonna is making no money.
Or Drake is making no money, but the guy who sings 'Happy Birthday' 10,000 times a week and uploads it through a DIY distribution system is now making more money than Drake and Madonna because this person has uploaded 100,000 songs. That is the path AI was on without attribution. Our attribution allows the specific tying of the outputs to the inputs so that the inputs that are more valuable to the process make more money. So, when something is generated, we can tell our clients, 'This input was 28% responsible for the output; this next input was 10% responsible.' Our clients have to use that to disperse funds to the artists they owe money to. Most rights holders, when they make money, have an obligation to share some of it with other people.
If they don't have attribution, they have no idea how to share that, and it's inherently unfair. So, we've made sure that the content that gets used and creates the most value in these systems receives the most remuneration, and I think that's how the business should work.
Are there any accepted "truths" about AI's role in music & creative industries that you disagree with? What's your contrasting viewpoint and evidence?
SEAN: I think there’s often this narrative about AI replacing humans—replacing musicians, composers, or people in general. But I like to think back to when the Internet first started. Looking back, we can see there were very similar fears at that time. We can point to industries that were wiped out by the Internet: music stores, maps, encyclopedias, fax machines, classified ads. These things just don't exist in the same way anymore. I want to acknowledge that this was very destructive for the people working in those industries. But, beyond that, a whole new set of industries emerged. In some cases, it made things better, and in others, it simply changed things.
Some might argue that certain industries are worse off now than they were before the Internet. But many would also agree that the Internet brought great net positives for humanity, even if there were some downsides too. It’s not black and white. AI will do something similar. It will wipe out some industries, and in ten or twenty years, we’ll look back and say, ‘Well, that industry is definitely gone.’ But the output of that industry will still exist, just in a different form. I think that’s what we’ll see happening.
However, the idea that AI is here to replace artists, like putting a quarter in a slot and getting a song, just doesn’t make sense. I haven’t talked to a single person who says that’s what they want from their music catalog. I’ve spoken to many people who talk about the emotional connection they have to an artist, to the artist’s story, to the stories told in their songs, and how these connect to the artists’ own backgrounds. I think AI will help more artists share those kinds of messages. No one wants to hear T-1000 croon to us like Barry Manilow.
How do you see AI in music evolving in the next 5-10 years? Where do you envision Musical AI in that future?
SEAN: I think AI is going to become much more integrated into the creative process. Just like we started integrating virtual tools and the ability to plug in acoustic, real-world instruments into our digital workstations, I believe AI will become a bigger part of the creative process with increasingly sophisticated tools. In the same way Photoshop is seen as a collaborator in an artist’s workflow, I think AI-infused tools will also be seen as collaborators rather than competitors.
All of this, I believe, needs to be built on systems like ours, where these advanced backend tools—like the ones that aren’t sexy but drive businesses, such as Adobe—will be provided. These systems will give artists and rights holders clear attribution and rights management to ensure fair compensation. In a world where AI plays a significant role in everything we do, it will be crucial to understand the influence of what AI is suggesting. Matt, anything else to add?
MATT: Oh, I wouldn't even try to predict anything about AI. I used to feel the same way about digital music—anyone who claimed they knew the future was either a fool or lying. The only thing I can guess about AI is that the term 'AI' will gradually fade away, and then one day we'll wake up, and it will be everywhere, just called by different names.
Can you share anything about the next thing you're working on?
MATT: We just launched our MVP. We've onboarded a ton of rights holders, and we're onboarding an AI company. I think the next thing we're really looking to do is scale for our AI company clients. We're working great with rights holders, and now our job is to help our AI company clients grow their businesses and excel at what they do. So, you won't see a big product release from us. What I hope you'll see—what I know you'll see—in the coming weeks and months are great products from other people that sit on top of our system, so they can differentiate based on their interface, their experience, and their training, and hopefully not have to worry about the boring stuff we take care of in the back end.
What advice would you give to aspiring entrepreneurs looking to enter the Gen AI space?
SEAN: I'll begin by saying that solving real-world problems with AI is crucial to succeeding in this space. There's plenty of room for innovation and for developing new tools that push the boundaries of AI. One of the keys to thriving here is understanding where you can provide genuine value, not just hype, and figuring out how to monetize that value. That's challenging, and not many AI companies have cracked that code yet.
MATT: My only comment would be: don't steal your training data.
Is there anything else you’d like to share about Musical AI or your personal journey?
SEAN: At our core is a profound respect for artists and their rights. We’re committed to ensuring AI grows in a way that works with humanity, not against it.
Browse the Music AI Archive to find AI Tools for your Music
Find out more about Musical AI
Musical AI Links: Website | LinkedIn | Sean’s LinkedIn | Matt’s LinkedIn | Nicolas’s LinkedIn