Spotify Will Label AI-Generated Artists and Pull Them From Default Recommendations

Spotify will start labeling AI-generated artist profiles in mid-September, and pull them out of the default listening surfaces that shape what hundreds of millions of subscribers actually hear.

Spotify announced the new "AI Persona" badge on Tuesday, according to its newsroom post. The label will appear in artist banners, "About" sections, search results, and next to AI-generated tracks in playlists. Clicking it shows whether the underlying creator self-disclosed the AI identity or was flagged by Spotify's own review.

The badge is the surface. The mechanism that actually changes the music economy is the distribution penalty that comes with it. Flagged profiles will not appear in Spotify's default editorial or algorithmic recommendations unless a listener already follows the account, Gizmodo reported the same day. Discover mixes, Radio, and similar surfaces will treat AI personas as opt-in. The label tells listeners what they are hearing; the distribution penalty decides whether they hear it at all.

The policy runs on two tracks. The carrot is self-disclosure: artists can flag their own AI use through Spotify for Artists and be marked transparently. The stick is a review process aimed at photorealistic AI identities, meaning convincing faces attached to people who do not exist, that begins with profiles that have crossed "defined audience thresholds." Spotify does not publish what those thresholds are, or how the review will identify which profiles cross them. That gap is the policy's biggest variable: how wide the review sweeps determines whether high-traffic AI personas are nudged into disclosing or simply cut out of default discovery.

Earlier in 2026, Spotify introduced "AI credits" that let producers disclose AI use in specific roles such as lyric writing or vocal production. The AI Persona badge extends that disclosure from production work to the artist's identity itself. A coalition of musician and label organizations proposed its own voluntary labeling system the prior month. Spotify's version is unilateral and carries an enforcement mechanism the industry's draft did not include. Whether the platform's approach will satisfy the coalition, or be dismissed as insufficient, is a live question for the rollout.

The backdrop is a regulatory void. The United States has no federal deepfake law, and AI artist personas, meaning fictional singers, producers, and bands that exist only as generated media, have proliferated across streaming services, blurring the line between novelty act and impersonation. Spotify is writing the rule because no one else is. The policy is not a ban on AI-generated music. It is disclosure paired with a distribution penalty, applied by a platform that controls the default listening surface for hundreds of millions of users.

The second-order effects cut in two directions. For human creators competing with AI personas in the same playlists, the policy is a partial moat. Their work remains default-distributable while AI competitors are pushed to the opt-in margins. For AI creators who disclose honestly, the label is a tax: transparency without protection, and the question is whether listeners will follow flagged accounts by choice. For listeners, the change will show up as fewer synthetic artists in their Discover feed and a clearer signal on the ones who remain.

What the policy does not yet say is how broadly the stick will land. Spotify calls the audience thresholds "defined" and declines to publish them, and the review process is described only in principle. Until those thresholds are visible, or until the first batch of flagged, undisclosed profiles is named, listeners and creators are working from the policy's intent rather than its reach.

The mid-September rollout will be the first concrete data point. The next one is what Spotify does with the review queue it has not yet described.