Not Fair – October 2025
Recently the discussion about paying royalties for playing music has become louder: Increasing possibilities of creating music with AI opens up the possibility to save a few bucks on the cost of playing music for some venues. On the musician side this has stirred up the discussion on the use of their music to train the algorithms that form the basis of the functioning of the AI models. If it wasn’t for the hard work of these musicians, there would not have been any AI music. Who is going to compensate them for all the work? But, on the other side, businesses and venues, operating in an already complex world of licenses and fees, face rising costs, increasing competition and rising administrative hurdles.
AI-generated music promises cost savings for businesses, but it raises ethical and legal questions. If AI models are trained on existing music—often without explicit permission—shouldn’t the original artists be compensated? After all, without their creativity, there would be no data to train these algorithms. Is the sound you want to use made by the artists that are used in the training data, or is this really something new? Composers also based their compositions on things they learned from others, and the way a musician plays is also influenced by the teachers or inspirations of other musicians. It this comparable to how AI works? Or not?
For decades, businesses have paid fees to artist rights organizations (like ASCAP, BMI, GEMA, BUMA/STEMRA, SENA) for the right to play music. These organizations then distribute royalties to musicians, composers, and producers. In theory, it’s a straightforward system, but in practice, it’s not that easy: you need to know which music is played where, but many venues don’t automatically report every song played. Moreover, many musicians and composers are also unknown to these organisations, and without accurate data, royalties can’t be distributed fairly.
Imagine you have to identify every musician in a gamelan group from Upper Kalimantan: this can be a monumental task. But even well-known artists may have early, unregistered work that might slip through the cracks. So, registration is key: If a musician or their work isn’t registered with a rights organization, they won’t receive royalties—no matter how often their music is played.
Another complicating matter: although music is a worldwide venture, each country still has their own different laws and institutions, even within the European Union. And on top of this, some musicians/authors do not want to be represented by a national organization, but use their own organization to deal with their artist rights. This fragmented situation means that for artist rights organizations, the retribution will never be fully fair and precise. Moreover, in current times with increasing economic protectionism, the flow of payments of artists rights and information about music plays between countries is meeting increasing difficulties.
As you can see, having a fair distribution of the authors rights is not an easy task. Nevertheless, it is important that we try: not only because we have the ethical duty to do so, but also to protect the value of music. Think of it like this: If we only plant monocultures, we lose the richness of the forest. If we only listen to AI-generated music trained on mainstream hits, we will lose the gamelan ensembles, the avant-garde jazz collectives, the folk singers from remote villages: all the things that might give way to new kinds of music. Music becomes a loop of the familiar, and will lose its power to surprise, to challenge, to connect us to cultures and emotions beyond our own.
Music360 