The record industry’s fight with Suno has a new theory of the case, and it targets how the company’s newest model was built.
Sony and Universal Music Group sued the music generator again this week, telling a court that v6 inherits unlicensed recordings from the models that came before it. Their complaint calls the practice “model laundering.”
In the labels’ words, training a new model on an infringing model’s outputs “does not eliminate the infringement; it launders it.” They describe v6 as “the fruit of the same poisoned tree,” arguing nothing changes unless Suno rebuilds from zero.
Method is the second front. According to Sony, Suno used distillation to shape v6 around the behavior of older teacher models.
Those teachers, the complaint says, were themselves built on infringing data.
Suno disputes the premise. Since launch the company has said v6 was “trained from the ground up, with a new set of data.”
Co-founder Jack Brody told The Verge the data set was new, without going into detail. Spokesperson Rachel Racusen later described the sourcing as licensed partner content, community creations, preference signals and the team’s accumulated learnings. Neither went as far as saying whether uploaded audio, or outputs derived from it, formed part of the mix.
Unlike several other rights holders, the two labels have signed no licensing deal with Suno. The filing extends a year of litigation and tests a doctrine that could reach any model retrained on its own synthetic output.
