music

A recent Senate policy brief warned that AI models could cut musicians’ incomes by as much as 25 percent within four years, as training sets devour the fruits of their labor without any licensing payments or bargaining leverage.

For these communities, hearing a generative model spit out a synthetic gospel track or an AI-generated country twang is not just an artistic affront, it can feel like outright blasphemy, proof that tech elites would rather manufacture culture than nurture it.

But labels and rights-holders see the same tools as built on an enormous act of industrial-scale theft. Decades of copyrighted recordings, the fruits of 100 years of human artistry, have been scrapped and repurposed as training data without so much as a licensing check or a royalty paid.