Brein Foundation shuts down artificial intelligence dataset

A large Dutch dataset that was used to train artificial intelligence (AI) models, has been taken offline by the Brein Foundation.

The Brein Foundation recently received a tip of a dataset that consists of illegal copies of tens of thousands of books, millions of articles from Dutch news outlets, and subtitles of countless movies and TV series originating from illegal sources.

This enormous dataset was compressed so it could be easily used for training purposes of large language models (LLMs). This is not permitted because it’s basically the same thing as data mining from illegal sources. Also, the copyright holders never gave their permission to use their works for training AI models.

“The Brein Foundation takes action against infringement and unlawful use of artificial intelligence. We’ve inspected the dataset for ‘Nothing from this publication may be reproduced’ and this yielded more than 10,000 results. Each of these concerned illegally copied books. The news articles were also copied from websites with copyright reservations. This clearly shows that copyrights have not been respected,” CEO Bastiaan van Ramshorst says in a statement.

The dataset has been taken offline. The owner signed a statement promising no longer to infringe copyrighted materials. Furthermore, he shared what companies received the dataset. The Brein Foundation is now investigating which AI models have used the dataset and will contact the parties involved.

The Brein Foundation is an advocacy group that represents the interests of musicians, actors, writers and other artists working in the Dutch entertainment industry.

The foundation has made a name for itself by launching numerous court proceedings against individuals and piracy websites for copyright infringement, most notably against The Pirate Bay, Library Genesis, Anna’s Archive shadow libraries, and Dutch hosting service provider XS Networks.

In November 2021, the Brein Foundation and the Dutch Copyright Federation have drawn up an agreement with six Dutch internet service providers (ISPs) and the Justice Department, in which the participants agreed to block internet domains that illegally offer copyrighted material if a court orders to do so. Ever since, the Brein Foundation is in charge of updating the domain list of blocked websites.