European Parliament Legal Affairs Committee adopts report calling for opt-out, transparency and remuneration for copyrighted works used to train generative AI

The European Parliament’s Legal Affairs Committee (JURI) adopted an own-initiative report setting out proposals to ensure transparency and fair remuneration for rightsholders when copyrighted works are used by generative artificial intelligence (genAI). The committee approved the proposals by 17 votes in favour, 3 against and 2 abstentions. MEPs want EU copyright law to apply to all genAI systems available on the EU market regardless of where training takes place, with rightsholders able to refuse the use of their content for AI training and automated data crawling. The report calls for AI providers and deployers to provide full transparency, including a list of each copyrighted work used and detailed records of crawling activities, and warns that failing to meet these requirements could be treated as copyright infringement with legal consequences. It also calls for fair remuneration, asks the European Commission to examine whether remuneration could apply to past use, and rejects a global licence model based on a flat-rate payment. Further proposals include protecting media pluralism by giving the news media sector control over use of its content for AI training, stating that content fully generated by AI should not be protected by copyright, and requiring digital service providers to act against illegal dissemination of manipulated or AI-generated content. The Commission is also invited to facilitate voluntary collective licensing agreements by sector and explore tools enabling rightsholders to prevent use by general-purpose AI systems. The report is due to be put to a Parliament plenary vote in March.