Publication details
Collective Licensing for Gen AI Training: Feasible or Flawed?
| Authors | |
|---|---|
| Year of publication | 2026 |
| Type | Article in Periodical |
| Magazine / Source | EUROPEAN INTELLECTUAL PROPERTY REVIEW |
| MU Faculty or unit | |
| Citation | |
| web | https://www.sweetandmaxwell.co.uk/en-gb/products/european-intellectual-property-review-journal-and-ejournal-subscription-30928131 |
| Keywords | artificial intelligence; collective licensing; collective management |
| Description | This article examines whether collective rights management—in voluntary, extended, or statutory forms—can provide a legally and technically viable framework forlicensing copyright-protected content used for generative-AI training in the European Union. The shortcomings of the CDSM Directive’s text-and-data-mining (TDM) exceptions and the limited effectiveness of opt-out mechanisms have triggered renewed debate on collective licensing as a means to ensure legal certainty and fair remuneration (such as statutory licenses, limitation-based right to remuneration, extended collective management or mandatory output-based remuneration falling on providers bringing fully developed AI systems on the market). The paper analyses each model’s feasibility, focusing on their legal justification, proportionality and technical realities. |