Music Industry

Musicians Are Losing Millions to Metadata Errors. Can AI Actually Fix It?

Credit: Outlever

Key Points

  • The music industry's AI debate focuses on creative replacement, but creators face a more immediate problem: losing hundreds of millions annually to metadata errors that prevent proper payment.
  • Chris McMurtry, EVP at OpenPlay and co-founder of ONCE.app, uses AI to automate registration across disconnected systems, eliminating the need to manually re-enter metadata across distributors, PROs, and collection agencies.
  • By solving the "representation problem," McMurtry's approach shows that AI's real value in music isn't replacing creativity—it's automating the complex administrative work that keeps creators from getting paid.
Chris McMurtry - EVP of Global Enterprise Innovation | OpenPlay
I was spending more time entering the same data in multiple places than I was composing the music itself.Chris McMurtry - EVP of Global Enterprise Innovation | OpenPlay

The dominant conversation about AI in the music industry centers on creation. High-stakes legal battles over copyright and a constant stream of publisher lawsuits dominate the headlines. But while many industry leaders are fixated on creative replacement and platforms race to build detection tools, one of AI’s most immediate impacts is happening at the systems layer: the unglamorous but foundational infrastructure that governs attribution, rights, and payment. It’s a pain point that burdens creators with redundant data entry across disconnected platforms when their time would be better spent creating.

Chris McMurtry lives a double life as a multi-platinum recording artist and composer who also builds enterprise-grade catalog and rights systems for major labels as EVP of Global Enterprise Innovation at OpenPlay. A former Apple technologist recognized as one of Billboard’s “2019 Digital Power Players,” he co-founded ONCE.app, an AI-powered platform that automates metadata registration across music industry databases. His work fits into a wider industry trend of tech investment in music discovery and metadata.

For many independent artists, the act of creation is followed by an overwhelming amount of data entry. For McMurtry, the breaking point wasn’t creative burnout, but administrative friction as his catalog grew past 700 works. McMurtry co-founded ONCE to tackle complexity head-on. This AI "registration agent" is built on a core principle of radical transparency to prevent the technology from becoming a black box. Artists can see, edit, and even delete what the AI remembers about them, retaining the final say.

  • Death by data entry: He found himself stuck in a loop, entering the same metadata across platforms including his distributor, ASCAP, SoundExchange, and the MLC. The volume became unsustainable. “I was spending more time entering the same data in multiple places than I was composing the music itself,” says McMurtry.
  • Representation, not replacement: For McMurtry, this reflects a fundamental reframing of where AI's real value lies. “The music industry doesn’t have a creation problem. It has a representation and systems problem,” he argues. “AI becomes powerful not when it replaces creators, but when it preserves their intent across broken, fragmented systems.”

The agent handles the context surrounding creativity, managing the key but often obscure data points that make sure an artist is properly credited and paid.

  • Show me the number: For McMurtry, the value becomes clear when you ask a simple question that most artists can’t answer. “I often challenge famous songwriters to name their IPI number. A system can only pay a songwriter if it is given that specific number. One of my friends, for example, doesn't know her IPI number, even though that is how she gets paid. An AI agent solves this by doing the work for them. It infers the information, sees the artist's name, connects it to their work, and pulls the IPI number needed for payment.”
  • Too fast to trust: Shaped by his time at Apple, McMurtry’s design philosophy prioritizes making complexity feel simple. But in early tests, ONCE’s efficiency created an unexpected problem in early testing.“The first version of ONCE was so fast, I had to deliberately slow it down. We could get a song from finished to delivered in fourteen seconds, and users thought it was broken because they didn't realize the app had already inferred all the necessary information,” says McMurtry. “I had to make it ask one question at a time, even though it already knew the answer, simply because users weren't psychologically ready for something to be that easy.” The adjustment proved essential: when AI eliminates friction too completely, users need artificial slowness to trust that the system is working.

While this practical approach is finding an audience with everyday creators, it also highlights a clear divide within the music community. While there is a growing acceptance of partially AI-assisted works, McMurtry sometimes faces resistance from within his own company.

  • A house divided: The split reflects different vantage points within the industry. “I get pushback, even from my own co-founder, a 16-time Grammy-winning engineer,” says McMurtry. “He has been very afraid to be AI-forward, arguing that musicians hate AI. But he’s talking to a specific tier of successful artists, like a Keith Urban, who is not the intended user of a tool like ONCE. When he returned from the NAMM conference convinced that everyone hates AI, I had to point out that the opinions of people who can afford to fly to an industry conference don't represent the common musician.”

Beyond artist sentiment, McMurtry points to a deeper issue. The most vocal industry opposition to the proliferation of new music is rooted in a flawed economic model that benefits incumbents. He directly refutes the AI slop argument by explaining the financial incentive behind it.

  • Follow the money: McMurtry sees this resistance as more than artistic pride. He points to a deeper structural issue driving industry opposition to new music. “The industry argument about 'AI slop' is a red herring,” McMurtry contends. “The real issue is the pro-rata payment model. As more music enters the system, the pool of money gets divided among more tracks, which reduces the market share for major labels and, therefore, their portion of unclaimed royalties. It’s a flawed system. We know who listens to what. The royalty money should go to the artists you actually listen to. Anything else is like going to Whole Foods, buying bananas from a local grower, and having all the money go to Dole just because they have the market share.”

McMurtry’s mission is to demystify the technology by showing musicians they are likely already using it. Familiar tools like Moises for stem separation are AI-powered, even if users don’t think of them that way.

  • A friendly front door: “I see these tools as a front door into the world of AI,” he explains. “For musicians who are up in arms, I can show them they may already be using AI to distribute their music. Once they see that it's not all bad and can be a practical help, the fear starts to fall away.” By creating a positive first experience with administrative AI, ONCE becomes a gateway rather than a threat.

McMurtry reframes artificial intelligence as a practical tool, rather than a creative threat. For him, AI's immediate power in music is its ability to enforce order and consistency in a chaotic digital world. His goal isn’t to build another company, but to solve a fundamental problem for all creators. “AI isn't magic. It simply follows instructions,” he says. “The reason it works so well for metadata and operations is because it provides continuity and ensures consistency.”