A look at YouTube's negotiations with record labels over an AI tool that lets creators perform using major musicians' voices, a pivotal moment for AI in music
Lucas Shaw / Bloomberg : X: @lucas_shaw . LinkedIn: Ben Liebmann and Steven Wolfe Pereira See also Mediagazer X: Lucas Shaw / @lucas_shaw : US customers cancelled streaming services at record rates last month. Rates of churn at Disney+, Hulu, Apple TV+ and Max are all way up, per @AntennaData https://www.bloomberg.com/... LinkedIn: Ben Liebmann : Music has been at the frontlines of almost every major change in the consumer internet. It was one of the first industries decimated by piracy … Steven Wolfe Pereira : We are at the dawn of the #AI era - and we haven't seen anything yet. #ChatGPT is the equivalent of AOL Messenger- we're just getting started in seeing new applications and use cases for AI. … See also Mediagazer
Context & Ripple Effects
The talks extend YouTube and Universal Music Group's earlier partnership to explore AI music products and artist payment, and follow reports that YouTube was developing a creator-facing voice tool and seeking training rights from major labels for that planned product.
They matter because the product’s commercial path depends on converting artist voices from an unlicensed generative-AI input into a negotiated, compensated use case on a major distribution platform.
First-order effects
- YouTube and the major labels must settle permissions for voice use, training, and artist compensation before the tool can be broadly deployed with major musicians’ voices.
- Artists and labels gain a direct negotiating role over how recognizable voices are made available to creators, rather than leaving that use solely to platform policy.
Second-order effects
- Any terms reached would give other AI-music products a concrete licensing and payout benchmark to contend with; subsequent reporting of talks with Sony, Warner, and Universal shows the same rights holders becoming central gatekeepers.
- Creator tools built around licensed voices could favor platforms able to secure label relationships, while unlicensed voice-generation services face a sharper distinction from authorized offerings.
Third-order effects
- If such agreements become repeatable, generative music is likely to move toward a rights-managed model in which labels license catalogs and identities while platforms handle creator distribution and monetization.
- The longer-term fault line will be whether licensing terms can accommodate broad creator use without eroding artists’ control or making access exclusive to the largest platforms.
The trend: This is one early point in the commercialization of generative music, where AI voice capabilities are being routed through established music-rights licensing rather than treated as a purely open creation tool.