ISTANBUL

Dutch journalists’ and writers’ unions on Friday formally demanded that Meta cease using copyrighted works by Dutch authors, reporters, and translators to train its artificial intelligence models.

The Dutch Association of Journalists (NVJ), the Authors’ Union (Auteursbond), and the writers’ rights organization Lira said Meta, the parent company of Facebook and Instagram, trained its Llama AI language model on texts obtained from illegal datasets without consent or payment.

In the demand letter, the unions urged the US tech giant to immediately stop using the disputed material and halt offering AI models trained on such data, the Dutch news broadcaster NOS reported.

If Meta does not respond to the letter, a summons is expected to follow, said NVJ chairman Thomas Bruning.

The dispute stems from US court documents indicating that Meta allegedly downloaded tens of terabytes of text from an illegal online database containing books, articles, and other copyrighted works, including those by Dutch authors.

The unions argue that copying and distributing such texts without consent violates copyright law and undermines both the economic and creative positions of writers.

“Meta is pursuing significant commercial interests with these AI models and expects substantial economic returns,” the organizations said, warning that creators risk being sidelined by systems trained on their unpaid work.

Liesbet van Zoonen of the Authors’ Union said the AI industry is worth billions and should compensate writers, translators, and journalists for the use of their content. “Authors are not a free resource for AI,” she said.

Bruning added that other AI companies could face similar actions in the future, stressing that “we have to start somewhere.”