ElevenLabs just dropped a new model that can generate full-blown music tracks, and yes, according to them, it is all cleared for commercial use. This marks a big leap beyond their usual territory. Until now, ElevenLabs has made a name for itself with eerily realistic text-to-speech tools, chatbots, and speech translators. But now, it’s officially entered the AI music game.
To show off, the company released a few samples, one of which features a synthetic voice rapping about hustling “from Compton to the Cosmos.” It’s a clear nod to the lyrical legacy of Dr. Dre and Kendrick Lamar, but unlike the OGs, this AI never lived it. And that’s exactly what’s making people uncomfortable.
The tension? Generative music AI has a copyright problem. Just last year, music startups Suno and Udio were slapped with lawsuits from the Recording Industry Association of America (RIAA), accusing them of training their AI on copyrighted tracks without permission. Now, both companies are reportedly trying to work out licensing deals with the major labels.
To sidestep those legal landmines, ElevenLabs has struck partnerships with Merlin Network and Kobalt Music Group, two big players that license music from independent artists. Merlin’s lineup includes names like Adele, Mitski, and Nirvana. Kobalt’s roster? Beck, Bon Iver, Childish Gambino, you get the idea.
Kobalt made it clear: artists have to opt in before their music is used to train the AI. The deal includes revenue-sharing, protection against misuse, and terms that match what most major publishers would expect. Licensing is one piece of the puzzle. AI can remix culture, but should it? Who owns the vibe when machines spit bars from borrowed pain?
