How AI-Generated Music Is Forcing Labels to Rewrite Their Contracts

Record labels spent the last two years fighting AI-generated music in court. Now they’re quietly negotiating licensing deals with the same technology — while rewriting artist contracts to cover a threat nobody anticipated when those contracts were first signed.

AI music generators capable of producing a convincing song, vocal included, in a specific artist’s style have gone from novelty to genuine commercial threat in a short span of time. Several viral tracks built from cloned vocals have racked up streaming numbers rivaling actual releases, forcing platforms to pull them down under likeness and copyright claims after the fact rather than before.

Major labels have responded on two fronts at once. Legally, they’ve pursued the AI companies training on their catalogs without a license. Commercially, several have simultaneously begun negotiating their own licensing agreements with those same companies, betting that controlled, compensated AI tools are preferable to an uncontrolled black market of unlicensed clones they can’t fully stop anyway.

New contracts, new clauses

The more lasting change may be happening in the fine print. Newer artist contracts increasingly include explicit language governing vocal likeness, training-data consent, and revenue splits for any AI-assisted material, terms that simply didn’t exist in standard boilerplate a few years ago.

“Every contract we write now has to answer a question nobody asked five years ago: what happens if a machine learns to sound like you?”

Artists themselves remain split between treating AI tools as a new instrument to experiment with and viewing them as an existential threat to the value of a distinctive voice. Either way, the industry consensus is that unlicensed vocal cloning isn’t going away on its own, which is pushing licensing deals from a defensive afterthought into a core part of how labels plan to make money from the technology rather than simply fight it.