It’s getting harder and harder to ignore generative music, with online discourses frequently disputing the validity of tracks and popular AI songs controversially making their way into mainstream listening. Earlier this year, Apple Music told Billboard that more than a third of the music submitted each month is entirely AI-generated. Middle Eastern music streamer Anghami told WIRED Middle East that AI submissions have doubled since 2023, averaging at around six million each month, while Deezer reported that fully AI-generated tracks averaged 90,000 uploads a day in June.
But increasingly, the challenge is perhaps no longer just about keeping AI music out. Rather, it is deciding which forms of it deserve a place. SoundCloud now works with more than 10 approved AI partners. Under its agreement with ElevenMusic, Pro subscribers will get three months of ElevenMusic Pro free, which allows up to 1,000 song creations a month. Tracks published that way will carry a disclosure naming ElevenMusic as their source.
Yet, if streaming platforms are concerned about the sheer volume of AI entering their catalogs, how much more of it are they prepared to encourage themselves?
Who Carries the Burden of Proof
And for all the industry’s assurances that human artistry remains at its centre, the question of accountability is becoming increasingly harder to sidestep. Much of the responsibility for declaring the origins of synthetic music is falling on the very people uploading and distributing it.
Labelling systems happen to bite the hardest on synthetic catalogues so far. Spotify’s AI Credits lets artists declare which roles and instruments were made with AI, which is then visible to listeners. The platform says it is also working on even more ways for listeners to quickly and easily see when an artist has disclosed AI use. It also launched AI Persona badges applied to artist profiles whose identity Spotify believes to be AI-generated.
Early this year, Apple added AI Transparency Tags to the specifications that labels and distributors use when delivering music. By the end of this year, those labels will be visible to the public as “Made With AI”. Notably, these labels are declared by whoever delivers the music, with the platform itself so far not doing any independent checks of its own.
The system, in other words, relies on a whole deal of honesty, and the distinction between transparency and trust becomes a rather important one. Anghami has no listener-facing AI labels yet, and says it’s a matter of caution rather than reluctance. “We would rather take the time to get this right than rush something out that wrongly flags a real artist,” says Mohammed Al Ogaily, COO of Anghami. Verified artist profiles have existed on the platform since 2017, and, Al Ogaily says, “these verified badges allow users to identify the real artists and ensure that no fake releases land on their profiles.”
Yet, even if AI in music becomes responsible, who stands to make money from it? For now, the industry’s position remains something of a balancing act.