AI spam tracks uploaded to profiles of deceased jazz musicians

Among the affected musicians are saxophonist William “Sonny” Criss, vocalist Jean Carne, cornetist Red Nichols and clarinettist Jimmie Noone. New releases have appeared on their profiles alongside AI-generated artwork, with several attributed to a publisher listed as “Ami bacuk”.

Criss, who died in 1977, had two new tracks listed on his Spotify profile. The songs featured anime-style artwork and were credited to a composer identified as “Zainul Irpan”.

The tracks appear to be newly generated jazz-influenced recordings rather than reworkings of identifiable recordings from Criss or another established artist. Their appearance on the profile of a musician who has been dead for decades has raised questions about how such releases were attributed to his catalogue.

Noone, a Black Dixieland clarinettist and bandleader who died in 1944, also had four singles added to his profile in April. ‘You Have Hurt Him’, ‘You Have Hurt Him 2’, ‘He’s Still There’ and ‘He’s Still There 2’ are pop ballads featuring what appears to be an AI-generated female vocalist.

The releases were accompanied by AI-generated artwork that differed markedly from the black-and-white imagery traditionally associated with Noone’s catalogue.

Carne, who remains active, has similarly had apparent AI-generated material added to her YouTube Music profile. The recordings resemble broadly produced jazz but do not reflect the distinctive vocal style associated with her work.

The cases highlight the difficulty streaming services face in preventing incorrectly attributed releases from appearing on established artist profiles, particularly as the volume of AI-generated music increases.

A Spotify spokesperson told Futurism that “the releases have been removed from the artist profiles” and that “protecting artists’ identities is a top priority for us, and we continue to invest heavily in detection and prevention.”

Spotify has also been testing its Artist Profile Protection feature, which allows artists to review and approve or reject incoming releases before they appear on their profiles.

The feature is part of the platform’s broader response to cases in which music is incorrectly uploaded or attributed to established artists. The incidents involving jazz musicians add to concerns about the ability of streaming platforms and distributors to verify artist identities and catalogue ownership before new material is published.

Other services are taking different approaches to AI-generated music. Deezer, for example, has developed systems to identify AI-generated tracks and exclude detected material from its royalty pool.

The growing use of generative AI in music is therefore creating challenges not only around copyright and royalties, but also around artist identity, catalogue integrity and the accuracy of information presented to listeners.