Without a specific agreement to the contrary, artists had agreed to a clause allowing their content to potentially be used for informing, training, developing, or contributing to AI technologies or services in the course of providing platform services.

But Seton says that “in the coming weeks,” that line will be replaced with this:

SoundCloud has now assured that it will not utilize artists’ content to train generative AI models designed to replicate or mimic their voice, music, or likeness without obtaining explicit consent, which will be obtained through an opt-in mechanism.

Seton reiterates that SoundCloud has never used member content to train AI, including large language models, for music creation or to mimic or replace members’ work. And, echoing what a SoundCloud spokesperson told The Verge in an email over the weekend, Seton says if the company does use generative AI, it “may make this opportunity available to our human artists with their explicit consent, via an opt-in mechanism.”

Ed Newton-Rex, the tech ethicist who first discovered the change, isn’t satisfied with the changes. In an X post, he says the tweaked language could still allow for “models trained on your work that might not directly replicate your style but that still compete with you in the market.” According to Rex, “If they actually want to address concerns, the change required is simple. It should just read “We will not use Your Content to train generative AI models without your explicit consent.”

SoundCloud did not immediately respond to The Verge’s request for comment.