A new article hosted on Penn State's JLIA site argues that compulsory licensing—rather than ad hoc lawsuits or opt-out schemes—could resolve the standoff between rights holders and AI developers over training data. The framing echoes a growing academic push to treat AI training as a licensing problem solvable through statutory or compulsory mechanisms, similar to how music sampling and cable retransmission disputes were eventually settled.
For data marketplaces and rights-clearance vendors, any momentum toward a compulsory scheme would mean standardized rates and reduced litigation risk, but also less negotiating leverage for individual publishers. Details of the proposed licensing structure are not specified in the available text.
The Implementation of Compulsory Licensing as a Solution for Copyright Conflict in Generative AI Training Data
