Spotify rolls out AI safeguards: impersonation, spam, disclosure

Spotify announced on Thursday a new trio of measures aimed at protecting artists and listeners from fraudulent and low quality content generated with artificial intelligence, saying the music streaming service removed 75 million spammy tracks over the past year. The company outlined updated impersonation rules, a music spam filter set to roll out in coming months, and voluntary AI disclosure fields tied to industry metadata standards in a blog post and a briefing with executives.

Spotify said the moves respond to a surge in generative AI tools that make it easier for bad actors and content farms to mass produce tracks, impersonate artists and exploit streaming mechanics to siphon royalties. Charlie Hellman, Spotify's global head of music product, said the company will not ban AI music outright. We are not here to punish artists for using AI authentically and responsibly, he said, adding Spotify hopes AI tools will expand creative possibilities.

Under the new impersonation policy, unauthorized AI voice cloning and deepfakes will be explicitly prohibited unless the impersonated artist has granted permission, the company said. Spotify also said it is increasing investment to stop...