Dutch authors and journalists demand Meta stop using copyrighted works for AI training
Dutch authors, journalists, and translators' unions sent a formal demand letter to Meta on 27 February 2026, requiring the company to stop using their copyrighted works for training the Llama AI model without permission or payment. The unions contend Meta obtained the texts from illegal datasets including Library Genesis, violating copyright law and undermining creators' economic and creative positions.
What happened
Dutch writer and journalist unions—the Auteursbond, Nederlandse Vereniging voor Journalisten (NVJ), and Lira—sent a formal demand letter on 27 February 2026 requiring Meta to stop using copyrighted Dutch texts for AI training.
Meta trained its AI-language model Llama using copyrighted works from Dutch writers, journalists, and translators obtained from illegal datasets including Library Genesis (LibGen), without permission or payment.
Reported impact
The unions contend that Meta's use of illegally-sourced copyrighted material undermines the economic and creative position of writers, translators, and journalists while Meta pursues significant commercial interests and substantial economic returns from AI models trained on such stolen works.
Index analysis
The unions stated that the AI industry represents a multi-billion-dollar business that now illegally uses and appropriates the work of writers, translators, and journalists without compensation, contrasting with creators being sidelined by systems trained on their stolen work.
Governance controls
The unions demand that Meta cease using illegal data sources such as LibGen and stop offering AI models trained using such sources in Europe, and that Meta establish a suitable licensing arrangement with creators.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.