Apple Music will soon label songs made with AI
Spotify was the first one to act, saying that they would label AI-created artist profiles
Apple Music is preparing to make AI involvement in songs visible to every listener, not just industry insiders. According to an email sent to content partners and obtained by The Hollywood Reporter, the streaming service will roll out "Made With AI" labels later this year, though no exact launch date has been set.
This follows an initiative launched by Apple back in March, where it allowed record labels and distributors to optionally label any song "materially generated using AI". In that instance, labelling was optional and not mandatory.
With this latest development, Apple has made the process mandatory for all content providers. This means any content provider must put in place AI transparency tags when there is a material part of a track created using generative AI.
According to Apple, this means any song "primarily derived from a generative AI service".
According to Apple Music VP Oliver Schusser, one-third of all the music on the platform is AI-generated, but this music only makes up less than 0.5% of all listenings on Apple Music. This is the reason behind the importance of such labelling.
This is because with no label, there is no way for listeners to differentiate the AI-generated music from artist-generated music on the platform.
Several weeks before Apple made its announcement, the RIAA and IFPI urged streaming companies to handle AI-made content like they handled explicit content, putting up a clear warning about it.
Spotify was the first one to act, saying that they would label AI-created artist profiles and not recommend them in their editorial playlists.
Apple’s recent decision to increase fines for any streaming fraud several times over the past months also points in the same direction, according to Schusser.
Apple hasn't disclosed how the new labelling requirement will be enforced, or what happens if a distributor mislabels a track. Self-reported systems like this one rely on the honesty of the same labels and distributors submitting the content, which leaves an obvious gap for anyone using AI music specifically to manipulate streams or commit fraud.
