Source: Generative AI

Generative AI reports on recent US court decisions involving AI companies Anthropic and Meta regarding copyright issues. The courts ruled in favor of ‘fair use’ for AI model training on published works, provided the content was legally purchased. This means that if a user buys a book, for example, it can be used to train AI models without needing the author’s permission.

This outcome is a significant victory for AI companies, as it removes the possibility for authors and publishers to opt out of model training. The ruling emphasized the transformative nature of AI training on such content. However, the presence of pirated books in some training datasets remains a separate concern. These rulings provide important legal clarity for AI development and copyright compliance in the US.