AI can now help produce a polished song in seconds. The harder problem begins once that song exists: who owns the voice, who controls the rights, and who gets paid?
That tension shaped HumanX Amsterdam’s “Who Owns the Future of Music?” conversation between PRS for Music CEO Andrea Czapary Martin, EMPIRE founder and CEO Ghazi Shami, and moderator Jennifer Strong. Strong opened with the question hanging over the entire music industry: when AI can generate radio-ready music almost instantly, what remains uniquely human?
Czapary Martin’s answer was immediate: “Definitely not the models.” For her, ownership still begins with the composers, writers and performers who bring emotion and identity to music. But she also made clear that PRS does not see AI only as something copyright holders must defend themselves against. The organization is already using the technology to identify music usage, trace international revenue and improve attribution in an industry where some collection processes remain fragmented or manual. The contradiction is striking: AI can complicate copyright while simultaneously helping creators recover money the existing system struggles to find.
Shami approached the same disruption from inside the studio. EMPIRE, he explained, has long operated as both a music and software company, using machine learning and AI across analytics, legal workflows, contract search and production. But his most revealing argument came from his own experience as a young producer. He remembered having musical ideas he could not fully realize because hiring guitarists, bassists, drummers and other session players was expensive. AI dramatically lowers that barrier.
His phrase for what happens next was simple: “Access creates culture.”
Yet Shami immediately placed a limit on that optimism. “Music is about taste.” Giving millions of people powerful creative tools does not suddenly give everyone the judgment required to make something audiences care about. AI can broaden access to production, but access and artistic value are not the same thing.
Czapary Martin pushed that argument further. If AI creates an unprecedented abundance of music, she suggested, scarcity may actually become more valuable. She said live revenues for PRS members had more than doubled over the previous three years as audiences continued going to concerts. Her point was not that synthetic music will disappear, but that a recognizable human artist, live performance and emotional connection could become stronger signals of value when generated content becomes plentiful.
Strong then moved the discussion from creating music to recreating identity: what happens when somebody clones an artist’s voice?
Shami argued that the answer should center on artist choice. He said EMPIRE has been working with YouTube and Google teams around an approach he called Voice ID, which could allow rights holders to choose whether a synthetic use of an artist’s voice should be blocked or monetized. He also recalled an earlier case in which a songwriter used cloned voices to make demos resemble major artists before some of the material was released publicly, showing how quickly a useful creative experiment can become a rights problem once permission disappears.
The final exchange revealed that AI-generated music is only part of a much older problem: the infrastructure of getting creators paid. Czapary Martin said some digital payments can still take many months to reach writers and composers, questioning why a digital industry cannot move money faster when usage and ownership information are increasingly available electronically. Her priorities were clear: reduce the lag, increase transparency and direct more value toward creators.
Shami then disclosed that EMPIRE had been building Supply Chain, a rights-management system designed to connect areas including masters, publishing, merchandise and other music assets. He said a formal announcement was planned for ADE the following month, with the company set to be based in Amsterdam. At the time of the HumanX session, that remained a planned launch rather than a completed one.
HumanX therefore left the music industry with a question much bigger than whether AI can write a song. It already can.
The real competition now is over authorship, permission, identity, attribution and payment.
Because as generating music becomes easier, proving who owns its value becomes much more important.