When you buy through affiliate links in our content, we may earn a commission at no extra cost to you. Learn 
BREIN, a copyright enforcement group based in the Netherlands, has successfully taken down a large language dataset that was being offered for use in training AI models.
A Reuters report revealed that the dataset contained unauthorized data from tens of thousands of books, news sites, and Dutch language subtitles harvested from “countless” films and TV series without permission. This move amplifies ongoing debates about data usage in AI training.
According to BREIN Director Bastiaan van Ramshorst, the extent of the dataset’s usage is unclear, but the forthcoming EU AI Act is expected to bring more transparency to the industry. This new regulation will require AI companies in Europe to disclose the datasets used to train their AI models, shedding light on previously unclear data practices.
The dataset’s removal sparks discussions about copyright and the exclusive right to reproduce data, particularly in the context of AI/ML training.
Meanwhile, nine EU countries have filed a complaint against social media platform X for using posts without permission to train its Grok AI, highlighting the growing concern over data usage in AI development.
Some Legal Precedents and Ethical Considerations
The issue of unauthorized data usage has led to several high-profile legal battles. Companies involved in AI development have faced lawsuits for using copyrighted materials without permission.
Google was sued for allegedly using copyrighted content in AI training. OpenAI and Anthropic have been criticized for unauthorized and aggressive web scraping, sparking calls for transparency and raising concerns about their data collection methods. Similarly, Meta has been forced to halt AI operations in certain regions due to the unauthorized use of user content to train its models, highlighting the need for responsible data handling in AI development.
These cases underscore the industry’s struggle with data usage ethics and legality. For many of these Big Tech firms, signing licensing deals could be a way to circumvent these issues, train its models, and ramp up adoption. AI startups like OpenAI have signed multiple licensing agreements with content publishers like News Corp and Vox Media in a bid to train its LLMs on articles and intellectual property produced and owned by these brands.
As AI technology continues to advance, addressing these legal and ethical concerns will be crucial. The industry must adapt to evolving regulations and ensure that data usage practices align with copyright laws and ethical standards.
