It was through music that the market economy began, that the remuneration of works of the mind was imagined, and that industrial society, the mass production of objects and the extreme division of labor were heralded. More recently, the digital economy also began with music, where musicians were remunerated through subscriptions to music platforms, after having been remunerated by the lords who commissioned them, then by the bourgeois who bought concert tickets, then by record companies who paid them royalties.
Today, AI is turning all that on its head.
It enables platforms to produce music without human composers, using the immense catalog of existing works to train themselves. For example, the virtual band The Velvet Sundown has passed the million mark in monthly listeners on Spotify; and the video clip for Heart on My Sleeve, composed by a certain Ghostwriter977, racked up millions of views before it was discovered that it was made entirely of AI-generated vocals copying two Canadian bands. We’re also starting to see totally artificial DJs, able to host a party like a human, with their speeches and playlists; and to use AI to write soundtracks for films, to imitate artists’ voices, timbres and styles. In short, to generate low-cost pieces of music that can be used just about anywhere as background music.
The evolution is dizzying: music sales generated entirely or partly by AI will increase by 40% by 2025, and will soon represent almost 20% of music industry revenues, without music creators receiving their due: According to a report by the International Confederation of Societies of Authors and Composers (CISAC), in the absence of an effective global framework, by 2028 music creators will be deprived of a quarter of what they should be receiving as a result of the integration of their creations into AI-composed works.
To protect artists’ rights, still thought of as copyrights, legislation is beginning to be put in place providing for remuneration for authors/composers whose works have been used to train AI models. The European AI Act, soon to come into force, requires creators of AI music, and more broadly of any form of generative AI, to be transparent about their sources. Several European projects are exploring blockchain and watermarking solutions to automatically identify the use of a musical work in an AI or derivative production, and automatically pay out micro-royalties. I believe that these protections will be illusory, and that tomorrow’s artists will be remunerated in ways other than through copyright.
In this new economy, indeed, everything will change: conversational AI will enable artists, without the need for a record company, to compose, arrange, mix, master, produce a music video, adapt their works to video games, yoga, interactive advertising, create their own label, their own studio, their own marketing agency, retain control of their data and sound identity, maintain a personalized dialogue with their fans, offer them customized musical experiences. They can become the entrepreneurs of their works, even if they will still need platforms for distribution, visibility and monetization for some time to come.
Feeling their replacement by AI algorithms coming too, music distribution platforms (such as Spotify, Deezer, Apple music) are trying to get ahead of the game by allying themselves with record companies, which are themselves in great danger: for example, Spotify has signed an agreement with three major record companies promising artists to use AI “with and for them”, and guaranteeing them transparency, consent, remuneration and protection against cloned voices. They won’t succeed in keeping this promise, especially as the artist remuneration mechanisms provided for in these agreements will be largely illusory.
So, for artists, if they’re not careful, the upheaval introduced by AI will just be a change of master: after the feudal lord, the bourgeois and the record companies, will come the algorithms. Musicians will see their copyrights disappear, and will become mere employees, if not slaves, of the algorithms. They can escape this by becoming entrepreneurs of their creations, using the formidable potential of AI themselves and taking power in the organization of concerts.
While consumers, who are also threatened with becoming mere slaves to algorithms that tell them what to listen to and how, could become co-composers of what they listen to, imposing their choices, determining the form to be given to the work they have chosen to listen to and, like the artist, privileging the true, direct, living, irreplaceable exchange of the concert, a challenge to solitude and artificialization.
The only true freedom, once again, in music as elsewhere, is to create and control the fruit of one’s creation. AI can abolish this freedom, or multiply it.
Image:generated by AI.