In early July, The Velvet Sundown, a band attracting over 1 million monthly listeners on Spotify, the music, podcast, and video streaming service, admitted to being a product of AI. In their bio, the group wrote, "The Velvet Sundown is a synthetic music project creatively directed by humans and composed, voiced, and visualized with the assistance of artificial intelligence… Not quite human. Not quite machine. The Velvet Sundown lives somewhere in between,".

Previously, since June, The Velvet Sundown emerged as a phenomenon with a 1960s rock blended with indie pop style. The group released two albums, Floating On Echoes and Dust And Silence, which entered charts in several countries.

Promotional image of The Velvet Sundown. Photo: The Velvet Sundown

Promotional image of The Velvet Sundown. Photo: The Velvet Sundown

However, their complete absence from live performances and interviews, combined with AI-generated promotional images, sparked suspicion from the online community and music experts. When initial accusations surfaced, the group denied them, calling them "baseless,".

After their admission, the band repositioned the project – not as a "hoax" but as an artistic experiment. Their bio also states, "This is not a hoax - it's a mirror. An ongoing art provocation challenging the boundaries of copyright, identity, and the future of music in the age of AI,". The group continues to be active and released the album Paper Sun Rebellion on 14/7.

The song "Dust on the Wind" has over 1 million streams, the highest on the group's Spotify channel. Video: The Velvet Sundown

The emergence of this band has ignited many debates. Speaking to the Guardian, experts like Roberto Neri, CEO of the Ivors Academy, and Sophie Jones, Chief Strategy Officer of the British Phonographic Industry (BPI), proposed mandatory clear labeling for AI-generated content. They argue that consumers have the right to know the origin of the product they are listening to. The lack of transparency increases concerns about consent and trust.

In addition, copyright issues and artists' livelihoods are also a concern. Liz Pelly, author of Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist, points out that AI models can be trained on databases of works without the creators' consent or payment. This creates unfair competition and the risk of unauthorized exploitation of artists' intellectual property, especially for independent artists who are already struggling in the streaming economy. She emphasizes, "We need to ensure that not just the stars have their rights protected, but all artists, even those with smaller followings, have the right to know if their work is being used illegally,".

In response to these reactions, Deezer, an online music platform, has taken measures to address "virtual music". They have developed software to detect and label AI-generated products. Aurelien Herault, Chief Innovation Officer, says, "Right now, in the process of 'normalizing AI', I think platforms need to be transparent and inform users about what they’re listening to,". However, Herault also suggests that labeling may not be necessary in the future if AI music becomes widespread and musicians use it as a standard tool.

Spotify has a different approach. A spokesperson for the platform stated that they are merely an intermediary and the primary responsibility lies with the content providers who upload the music. This stance has faced criticism for being a way of shirking responsibility. Many worry that this could repeat past mistakes of the streaming industry, where large tech companies profit while creators are left behind, without proper protection for their work.

Khanh Linh (Guardian)