Spotify is rolling out an "AI Persona" badge for artist profiles whose public identity appears to be generated by artificial intelligence, and it keeps music from those profiles out of its editorial and algorithmic recommendations unless a listener follows the artist. The policy, announced on 11 August for launch in mid September, arrives as Deezer reports that fully AI generated tracks now make up more than half of the new music delivered to it every day.

What the badge says, and where it appears

The badge "signals to listeners that an artist's identity may be AI-generated and does not represent a real person," Spotify said in its announcement. It appears on the artist profile, in Search and on track rows across playlists. There are two versions, according to Spotify's help page for artists: "AI Persona" when the artist disclosed it, and "Likely AI Persona" when Spotify's own review reached that conclusion.

The real consequence is in discovery. "By default, Spotify will not include AI Personas in any editorial or algorithmic recommendations," the company wrote. Badged music stays searchable, but it reaches a listener's personal recommendations only if that listener "intentionally engages with the artist, for example by following them," the help page says. On 30 September the page said the badges were "gradually rolling out in the Spotify app".

Who gets labelled, and how to appeal

Artists have been able to disclose through Spotify for Artists since 11 August, but Spotify says it "won't rely on self-disclosure alone". Its review team looks for profiles whose name and imagery appear to represent "photorealistic AI-generated identities", starting with artists who have met "defined audience thresholds". Every badge Spotify applies is checked by a person first, Deadline reported. Flagged artists are notified and can confirm the badge or appeal; if they do nothing, it stays. Listeners are due to get a way to report suspected profiles in the coming months.

The line Spotify draws is identity, not method. "This badge is about the artist's public identity, not about how the music was made," the company said. A real person, a stage name or a fictional character that does not depict a photorealistic AI generated human gets no badge, even if the music is made with AI tools; how a track was made is covered by separate features, AI Credits and SongDNA, TechCrunch noted.

Deezer: half the uploads, a sliver of the listening

On 21 July, Deezer said that tracks it detects as fully AI generated had for the first time passed half of all new music delivered to it each day, at a June 2026 peak of about 90,000 tracks a day. In January 2025 the figure was about 10,000 tracks a day, or 10 percent of uploads, and in April 2026 it was 44 percent, TechCrunch noted from Deezer's earlier reports.

More than 50%
Share of new daily uploads to Deezer that were fully AI generated at the June 2026 peak, about 90,000 tracks a day (Deezer, 21 July 2026)

Listening tells a different story. Fully AI generated music accounts for between 1 and 3 percent of streams on Deezer, because the service keeps detected tracks out of algorithmic recommendations and editorial playlists. Much of that listening is not real: up to 85 percent of streams on fully AI generated tracks were identified as fraudulent in 2025 and demonetised, compared with 8 percent across the whole catalogue, the company says. Deezer will now take down AI tracks used for streaming fraud and those not streamed for at least six months.

We have reached a pivotal moment, but we have the technology to keep track of the development, and an ambition to do what's right.
Alexis Lanternier, CEO of Deezer, 21 July 2026

An industry drawing lines

In September 2025 Spotify said it had removed more than 75 million "spammy" tracks over the previous 12 months, and in April 2026 it introduced a "Verified by Spotify" badge for human artists, Music Business Worldwide reported. The case that made the problem visible was The Velvet Sundown, an act that gathered more than a million monthly listeners on Spotify in 2025 and described itself as a "synthetic music project guided by human creative direction", Euronews recalled.

Labels have pushed too. "Fans don't want AI slop, and there's no justifiable reason that this content should be algorithmically served to audiences on streaming platforms or siphoning money from human artists," Universal Music Group chief Lucian Grainge said on the company's second quarter earnings call, as quoted by Digital Music News. Other services chose other routes: Bandcamp bans AI music, Tidal cut off monetisation for it and Apple Music has voluntary tagging, according to TechCrunch.

The European backdrop

Since 2 August 2026, Article 50 of the EU AI Act has required providers of AI systems that generate synthetic audio, images, video or text to mark the output in a machine readable format, and deployers who use AI to create deepfakes, audio included, to disclose them, according to a practical guide to Article 50. Generative systems already on the market before that date have until 2 December 2026 to add the machine readable marking, under a provisional agreement reached in May, the guide notes. The European Commission says the deepfake duty covers audio built to mimic a real artist's voice, MBW reported. Spotify's badge is its own company policy and asks a different question: whether an artist's name and face belong to a real person.

What changes for artists and listeners

For acts built around an invented, photorealistic singer, disclosure and followers now decide reach on Spotify. TechRound, a UK technology site, said the default exclusion puts such profiles "behind a participation wall" and set out the counterargument: singling them out "makes a value judgement about authenticity that the platform perhaps has no particular authority to make." The route that stays open, it noted: "Keep a real person as the public face, use AI as heavily as you like behind the scenes, and the badge doesn't apply."

For listeners, the badge meets a demand surveys have measured. In an Ipsos poll of 9,000 people in eight countries, commissioned by Deezer in November 2025, 80 percent agreed that fully AI generated music should be clearly labelled, and 97 percent could not tell it apart from human made music in a blind test.