Overall, the project addresses two competing societal objectives.

• Remuneration. AI tools are increasingly capable of generating compositions and simu-lating musical performances, potentially replacing traditional roles in the music industry at lower cost. GenAI models depend on music created by human artists for training, however, making it necessary to compensate those who contributed to GenAI development.

• Unbiasedness. To reflect a broad range of musical expressions, GenAI models require access to broad datasets, including copyrighted music. Without the support of artists and rightsholders, GenAI outputs will lack diversity, undermining its potential for use across the full spectrum of musical expression.

This project seeks to reconcile these objectives, crafting fair remuneration policies that will support a thriving and diverse GenAI soundscape. This balanced approach benefits society by protecting artists’ and rightsholders’ interests while ensuring that GenAI can represent a wide range of artistic styles.

This project contributes to the University of Amsterdam’s Human(e) AI research priority area and its focus on the methodological, legal and ethical challenges of AI. The project also aligns with the Citizens, Society and Artificial Intelligence (CiSAI) platform, bringing social science perspectives into the conversation on AI’s impact on society.