Herremans approaches AI music through the problem of measurable contribution: how can creators be recognised and remunerated when generative systems learn from immense, mixed repertoires? Her key intervention is to consider whether computational models might support, rather than merely threaten, attribution by estimating influence, similarity or training contribution. This does not resolve copyright by technical calculation, but it redefines metadata as a distributive infrastructure. Royalties depend upon identifying works, creators, uses and chains of derivation with sufficient precision to make payment possible. The conceptual operation therefore links machine learning to rights administration while retaining uncertainty about what influence means inside statistical systems. Her work bridges music information retrieval, intellectual property and platform economics, revealing that fairness requires both legal entitlement and technical legibility. Without robust identifiers, transparent datasets and auditable accounting, compensation becomes structurally impossible. AI’s capacity to analyse cultural patterns may thus be redirected toward the governance of provenance, provided that its metrics remain contestable and institutionally accountable.