Spotify Will Stop Recommending AI Personas by Default
Starting in mid-September, virtual artist identities will be labeled and removed from personalized and editorial recommendations unless listeners actively choose to follow them.
Spotify is drawing a new line in music discovery: not between songs made with and without artificial intelligence, but between a real artist and a fictional identity built by AI. Beginning in mid-September, artist profiles that represent an AI-created persona rather than a real person will receive an “AI Persona” label and will no longer be eligible for Spotify’s personalized or editorial recommendations by default.
The change does not amount to a ban on synthetic music. AI Persona profiles can remain on the service, and listeners can still search for them, open their pages, and play their catalogs. What changes is the default path by which an unfamiliar artist reaches an audience. Spotify’s recommendation systems and editors will no longer surface those profiles as if they were ordinary human artists without an explicit signal of interest from the listener.
That distinction matters because reports describing the policy sometimes use the word “subscription.” Spotify is not requiring a paid Premium subscription to hear or re-enable an AI Persona. A listener can actively follow the artist profile. Spotify’s own help materials say following artists helps improve recommendations, making the new policy closer to an opt-in system than a paywall.
Spotify is also separating artist identity from the technology used to make a track. A human musician who uses AI for lyrics, vocals, instrumentation, or production does not automatically become an AI Persona. The company already has an AI credits feature in beta that allows specific AI-generated contributions to be disclosed at the song level. That approach treats AI use as a spectrum rather than a binary label on an entire piece of music.
Creators will be able to self-identify a project as an AI Persona, but Spotify does not plan to rely only on voluntary disclosure. The company will review higher-audience profiles using human judgment alongside AI tools, including checks for photorealistic synthetic identities. In uncertain cases, a profile may receive a “Likely AI Persona” label. Artists will be able to appeal if they believe the classification is wrong.
The policy builds on Spotify’s broader authenticity program. In April, the company introduced “Verified by Spotify,” a badge for artist profiles that meet standards around sustained listener interest, platform standing, and signals of a real artist presence. Profiles that primarily represent AI-generated or AI-persona artists are not eligible for that verification, while real artists who responsibly use AI as part of their creative process can still qualify.
Spotify has been moving against other forms of AI-enabled abuse as well. Its rules allow removal of music that clones a real artist’s voice without permission. In 2025, the company said it had removed more than 75 million spammy tracks in the previous 12 months, describing mass uploads, duplicates, and other forms of “slop” as threats to listener experience and to the distribution of royalties.
The new recommendation rule is therefore less about whether computers can participate in music and more about whether software should quietly manufacture the appearance of a human performer. A synthetic act can still build a fan base. It just has to do so with its nature visible and with listeners making the first affirmative choice.
That may be the most consequential part of the change. In a streaming economy where discovery is often automated, Spotify is deciding that the question “Who is this artist?” should come before the algorithm decides “You might like them.”
