AI Music Labels Won't Fix What's Actually Broken, Says Former Sony Executive

The music industry’s latest proposal sounds practical on the surface: slap“AI-generated”and“AI-assisted”labels on streaming music. The RIAA, IFPI, Recording Academy, and SAG-AFTRA all agree it’s necessary. But according to Drew Thurlow, a former senior VP of A&R at Sony Music and current professor at Berklee College of Music, the entire approach is solving for the wrong problem. He argues that a generic label won’t accomplish what the industry actually needs to protect artists and listeners.

The core issue is that almost nobody listens to most music on streaming platforms anyway. In 2025, roughly 88% of the 253 million tracks available got fewer than 1,000 plays. Most AI-generated music falls into this same graveyard—available everywhere, heard by almost no one. Slapping a label on tracks that nobody’s streaming won’t protect artists or change listener behavior. Additionally, defining“AI-assisted”has become nearly impossible since almost every professional recording now involves machine learning tools. Pitch correction software, mastering algorithms, stem separation technology—these are all technically AI. A label that describes the modern recording process won’t provide meaningful transparency.

Instead of generic warnings, Thurlow contends the industry should focus on specific, solvable problems: preventing unauthorized voice cloning, removing fraudulent uploads, protecting artists from impersonation, and keeping fully AI-generated music out of the royalty pool. These actions would actually protect artists and listeners instead of creating the illusion of a solution. The parallels to the 1982 panic over synthesizers are hard to ignore—the industry freaked out then too, only to watch artists like Taylor Swift use digital tools successfully without any backlash. What do you think the industry should prioritize first when it comes to AI in music?